Part 1/3: How Can AI Help Quality Assurance and Compliance Stay Ahead of the Game?

Conversation analytics can significantly impact various aspects of your business, especially Compliance and Quality Assurance. Today, we’ll explore how AI-driven conversation analytics, such as Sentient Machines, can elevate your QA processes and how it seamlessly integrates with traditional QA scorecards.

Enhancing Compliance and Quality Assurance with AI

For Quality Assurance teams in regulated industries, ensuring that advisors remain compliant is critical. Non-compliance poses reputational risks and can lead to substantial fines. However, maintaining compliance requires heavy investment in advisor training and ongoing assessments, particularly in today’s complex regulatory environment.

If your current approach involves manually selecting a random sample of conversations and scoring them, you’re not alone—80% of companies do the same. But while this process is familiar, it limits your potential to learn and improve. The few companies using AI-driven conversation analytics gain access to insights that would otherwise be out of reach, setting them apart from competitors reliant on legacy systems or keyword-based speech analytics.

What Can AI-Driven Conversation Analytics Offer Today?

AI is adept at processing large datasets, identifying specific events, behaviors, emotions, and anomalies. At Sentient Machines, we define an “outlier” as a conversation that signals compliance risks, poor customer experience, or agent disengagement—factors that could lead to churn or high turnover. Outliers reveal valuable insights, allowing you to focus on conversations that have the potential to make a substantial difference.

Key Advantages:

1. AI-Assisted Scorecards: AI can align with traditional QA scorecards, answering questions that conversation analytics can address autonomously. This approach lets AI handle straightforward scoring, leaving your team free to focus on more complex insights.

2. Augmented Manual Scorecards: If your scorecard is too intricate for AI to complete independently, AI can still analyze 100% of interactions, producing actionable insights. This hybrid model combines human expertise with machine efficiency, delivering results that neither could achieve alone.

3. Transforming QA into QA Analytics: Traditional QA emphasizes performance monitoring, often requiring proportional increases in staff as customer interactions grow—an expensive, time-consuming approach. AI transforms QA into a data-rich analytics function, enabling teams to derive deeper insights and more nuanced metrics in less time.

4. Targeted Competency Training: AI simplifies competency training by automatically identifying relevant conversations. For example, if you need to assess how new hires are handling specific tasks, like setting up Direct Debits, AI can pinpoint these conversations, allowing you to review precisely what you need.

Integrating AI into Your QA Process

Here’s how our AI-driven conversation analytics can elevate your day-to-day QA activities:

- Focused Conversation Selection: Rather than relying on random samples, AI targets the most impactful conversations. If less than 1% of your customers are vulnerable, random selection will likely miss this group. AI’s capability to monitor 100% of interactions uncovers critical insights that would otherwise go unnoticed, identifying over 500 distinct events, behaviors, and emotions—including vulnerability flags and shifts in customer sentiment—to help prioritize reviews.

- AI-Driven QA Analytics Co-Pilot: Think of AI as a co-pilot for QA, ready to support or fully manage the scoring process. It can assist in completing scorecards, tracking performance, and ensuring quality standards are met. By handling complex data analysis, AI makes the QA process faster and more effective.

- Real-Time Alerts: AI can monitor rare or high-risk events, alerting you instantly via email when they occur. This proactive approach allows teams to respond swiftly to urgent situations.

Overcoming the Limitations of Traditional Speech Analytics

If previous experiences with speech analytics left you unimpressed, you’re not alone. Many systems on the market operate as glorified search engines, focusing on keywords rather than context. However, effective conversation analysis requires understanding not only *what* is said but also *how* and *why*. 

A truly AI-driven analytics system like Sentient Machines interprets emotions, behaviors, and conversational dynamics, adapting to specific business needs and reducing false positives. The result? More accurate, actionable insights than traditional keyword-based systems can offer.

A Strategic Advantage

By merging human expertise with AI’s analytical capabilities, companies can not only meet regulatory standards but also enhance team productivity and foster stronger customer relationships. 

Stay tuned for the next article in this series, where we’ll explore how AI can boost team performance, productivity, and well-being.

Staying Ahead with AI-Driven Conversation Analytics: A 3-Part Series

Financial companies, especially those serving customers directly, often ask us how our technology can help them tackle industry challenges. Today, regulatory compliance is a top priority for financial organisations, and it’s more complex than ever. Stricter regulations, a surge in vulnerable customers, and an increasingly remote workforce mean that companies can no longer afford to take a passive approach to monitoring customer interactions. Failure to act can lead to severe fines, business disruption, and lasting reputational damage.

Although most companies record thousands of customer interactions daily, 80% lack the tools to analyze them effectively, continuing to rely on labor-intensive, manual processes to ensure compliance and gauge customer satisfaction. This approach is not only costly but also impractical—only 1-4% of conversations ever get reviewed, and the insights obtained often come too late to be useful.

To address this need, we’ve created a 3-part series detailing how our AI-powered solution can analyse 100% of customer interactions, reducing compliance risk, streamlining quality assurance, and enhancing agent effectiveness. Here’s what each part of the series will cover:

Part 1: Compliance and Quality Assurance

Compliance should be a non-negotiable priority for every organisation—not only for ethical reasons but also because regulators are closely monitoring. Our AI solution allows companies to monitor 100% of interactions, vastly improving upon the typical 1-4% review rate at a fraction of the cost. By empowering quality assurance teams with AI-driven insights, we boost their efficiency by over 90%, ensuring continuous and thorough compliance.

Part 2: Enhanced Team Performance, Productivity, and Well-Being

Using AI paired with behavioral science, we’ve developed metrics that identify key behaviours and emotions influencing team productivity. By focusing on these drivers, our solution enhances employee development, supports learning, and promotes well-being, leading to a more effective and engaged workforce.

Part 3: Elevated Customer Experience

An engaged and well-supported team has a profound impact on customer service—more so than sporadic customer surveys. While gathering feedback is valuable, much insight can be gained from daily conversations. Our Customer Experience Index leverages AI to predict satisfaction, assessing each interaction to anticipate whether a customer would rate it positively or negatively. This proactive metric provides real-time insights into customer satisfaction, enabling organisations to act swiftly.

Up to Know good?

Spend on speech analysis is expected to grow to above $3.5b over the next year. It has increased the insight in customer experience greatly, but it has big limitations and is mostly old technology with new algorithms being glued on to enable AI.

One of the challenges of working with speech analysis is the nature of conversation. In culture “so bad” can mean “so good”. “Sick” can be a problem or the highest compliment and of course irony can change “more than satisfied” into a damning indictment of your customer experience. 

Culture plays a huge role in our understanding of what is said and not said. Americans say pants and the English say trousers. The world’s population of English as a second or third language speakers use conversational constructs from their native languages that can change meaning.

Sentient Machines is a next-gen technology. It provides conversation, interaction, and experience analysis that looks at “what is said”, identifies “what is not said” and provides insight into “what it means”. It is about context.

Without context, “I saw a man on a hill with a telescope” could mean either I had a telescope or maybe the man did, and who had the telescope could be the most important part of the conversation – context is everything.

Built around self-learning AI, our platform can detect experiential constructs. In simple terms, emotions, feelings, sentiments, and behaviours. Not just the difference between apathy and empathy. It doesn’t just rely on keywords to understand the situation. It can identify vulnerable customers in hundreds of common scenarios. It can understand the reason for a call.

But Sentient Machines has a few other tricks that make it different. Out of the box it starts to build an understanding of your conversations. It generates the topics your customers and employees are talking about into easy to understand and clickable journeys for you to explore.

Its Sunburst technology allows you to dive deeper into subjects and heading more quickly so any user can find what they are looking for in a few clicks. The UX is designed to make it easier for every users journey faster and more effective. From HR and Team Management to QA and compliance. Operations to Sales a couple of clicks and you can find what you want.

Because it is next gen, it learns from every decision you make generating a clear understanding of each stage in your customers and colleagues’ journeys. It customises itself to your organisation with every click.

But that’s not all - you can connect your telephony in hours, not days. Along with your reporting platforms, CRM and other tools and insight sources. Finally, the promise of a single lens and one version of the truth is a reality. Our military-grade and FCRS certification is built into everything we do.

Sentient Machines customers, including government departments and financial institutions, rate the platform so highly that once turned on no-one has ever switched it off.

So, what about the cost? Well, there are no licence fees, no expensive professional services, and no massive marketing programmes to pay for.

So, if you want to find out why Sentient Machines win so many awards and have so many happy customers get in touch and I can show you more about your customer experience in a day than you have seen before.

Morris Pentel, Customer Success Director

7 Ways AI Can Transform Your Consumer Duty Readiness

Compliance is changing. The Regulations are becoming more clearly focused and the penalties continue to become more severe. The new Consumer Duty creates new challenges for organisations that can be met with AI.

Now a significant part of the legal duties focus on the so-called soft and human interactions. This is already recognised by regulatory bodies like FCA, who have introduced Treating Customers Fairly guidelines, and continue to bring more focus on the consumer protection with the new Consumer Duty Principle. 

I’ve been frequently asked recently: What is the role of AI in making sure those principles are followed for the best possible outcomes?

Here are 7 ways that AI-driven conversational analytics can be used to help you prepare and navigate through this new change in regulation. 

Conversation Analysis in this context is using AI to understand what is said and what is not said on a phone call, in a live chat with an agent, or through any other channel. With the ever growing processing power available to AI applications, conversation analysis can understand the context of a conversation, identify vulnerabilities, confusion as well as other behavioural patterns, and more importantly do it at scale.

How AI Can Help You to Prepare?

What can you do today to guard your operation and be 100% ready for the Consumer Duty?

1. Clarity of Communication 

   Treating Customers Fairly already relies on clarity as a very important communication element in your customer interactions. The Consumer Duty requires that firms must give customers information that they can understand. This goes beyond a requirement to read scripts and provide information. And how do you measure this in an objective way? This is where an advanced AI system can help. An advanced conversational analytics platform like Sentient Machines can identify: How confused is your customer? Are they seeking for clarification because they are confused, or simply because they are inquiring about further information? Has the common ground been established between the two speakers?

2. Pre-emption of Complaints 

   It is not surprising that FCA have identified reduction of complaints going to the Financial Ombudsman Service as one way to measure the outcomes of the Consumer Duty Principle. Especially in relation to fees or charges or inappropriate product or service sales. In an ideal world, there are no complaints, and with the right AI approach, you can achieve exactly this - by preempting them. For the few that still come through, you could be alerted immediately and react promptly to potentially recover the customer relationship as you will have time to dedicate to it.

3. Focus on Customer Needs 

   Understand your customer needs in real time and measure the impact of changes you introduce immediately. In his recent speech, the CEO of FCA, Nikhil Rathi, mentioned an AI solution that was implemented by a major bank that could predict with 99% accuracy a customer’s bank balance in a year’s time. When customers were presented with this innovation, they did not want the product integrated into their banking app. So how do you ensure that you truly listen to your customers and their needs? Is your AI there for the right purpose? In the UK alone, there’s 2 million hours of customer interactions over the phone every single day. Listening to your customers, understanding what they like and dislike is critical in meeting the requirement to do with meeting the customer needs. An advanced conversational analytics platform can measure and quantify your customer preferences, and monitor emerging trends so that you can react while it still matters.

4. Track Customer Objectives

   While AI is still not at the point to perform extremely complex human tasks, such as a subtle judgement of whether customer financial objectives have been met or not under complicated circumstances, and where the concussion needs to be reached from combining multiple sources of information, it is certainly able to assist humans, or at least provide guidance in that direction. Similar to understanding customer needs, AI can easily scan whether or not you’ve discussed customer financial objectives, and raise red flags otherwise. It could also provide cues as to whether or not those objectives are met and raise red flags for outliers that humans would then have to verify.

5. Identification of Vulnerable Customers 

   The definition of vulnerability is the one that is constantly evolving and changing and the ability to recognise early signs can bring many advantages to firms. With AI, you can identify those early signs and reach out in a more personalised ways to really help nurture your customers. How is their health? How is their family situation? How is their stress level? 

6. Treatment of Vulnerable Customers

   Depending on your vulnerable customer situation, you might want to set up different frameworks to help support them. AI can not only help you identify vulnerabilities, but also, once the vulnerability is detected, it can scan and help advisors understand if they have followed required procedures, or if they could benefit from further training.

7.  Monitor Your Agent Well-being

   What does an advisor's wellbeing have to do with the actual customer? Everything. They are the face of your company. Imagine speaking all day with customers who are in a challenging life situation. Perhaps they’ve lost their job, or they’ve had sad news with regards to their health. Or their family. At the core of fair outcomes for your customer, is the stability of your own team who’s responsible for handling those conversations. Advanced conversational analytics platform from Sentient Machines can help with identifying those critical moments that could make all the difference.

Conclusion

Humans and AI working together are stronger than working independently. It is humanly impossible to review and listen to all customer interactions, yet alone accurately identify red flags at scale. AI on the other hand can scan and analyse an enormous number of conversations in no time, letting you spend your time only on things that create maximum impact. 

Advanced conversational analytics can work for you in the background, sending you alerts for emergencies, but also to turbo-charge your training and bring the competency of your team to the required level quickly. 

Consumer Duty might be the perfect bridge between enabling companies to fully embrace regulations while at the same time building healthy relationships with their customers. And AI, and conversational analytics in particular, has a lot to offer to make this bridge a massive opportunity to fully embrace the current technical revolution taking customer service to the next level.

Top 5 AI Supervillains

We at Sentient Machines often talk about how we like to use artificial intelligence (AI) to help people. Our leading-edge AI platform reads and tracks intent, language, and emotion throughout customer interactions – basically, what is being said and why it is being said. And usually, we use our blogs to show everyone the benefits of using AI. 

Today, however, we thought we would do something a little different; using movie villains to illustrate how AI can go wrong. 

AI has been a popular subject in sci-fi movies for decades, with artificial intelligence portrayed as both friend and foe on the silver screen, with its darker side giving audiences some of cinema's most iconic villains, with many of these films depicting a world where people have taken technology too far and explore the devastating consequences of manmade sentience. Many of these movies take a more nuanced approach to the topic, often portraying the villains as misunderstood victims of humanity's greed. 

So, with all this in mind, let's look at cinema's top 5 AI supervillains.

5. Ultron in The Avengers: Age of Ultron (2015)

 One of the most recent AI supervillains comes in the robotic form of Ultron. Originally designed to be a part of a peacekeeping program created by the acerbic Tony Stark and his bad-tempered friend Bruce Banner, Ultron went rogue, deeming humanity the greatest threat to peace on Earth and attempted to commit extinction-level genocide against them.  

The overarching theme of the Avengers sequel is this fictional conflict between biological beings and artificial intelligence. But how fictional is it? Is AI like Ultron possible? 

While we are a long way from "Artificial General Intelligence", which, contrarily from what is out there today, can think, feel, and reason as much as us humans do, Ultron embodies the "What Doesn't Kill Me Makes Me Stronger" ethos, which perfectly sums up AI and machine learning pretty nicely.

  4. Agent Smith and the Agents in The Matrix Quadrillogy (1999-2021)

 Agent Smith, played superbly by the terrifying Hugo Weaving, has become one of cinema's most iconic villains. With their CIA attire, dark sunglasses, and inhuman blandness, the Agents were AI characters whose singular purpose was to stop humans from discovering the truth about the reality of their universe. 

The original trilogy pit man against machine in a clearly drawn battle while also exposing humans as more machine-like than perhaps first thought and that machines are capable of possessing human qualities as well. Humans, for their part, are depicted as relentlessly driven as machines. Morpheus's faith in the Oracle's prophecy is unwavering, and his own followers follow him automatically. Trinity's loyalty to Neo has machine-like consistency. While Keanu Reeves' Neo exudes an almost robotic calm, and both he and Carrie-Anne Moss wear sleek, androgynous clothes. The Agents, by contrast, are fluid, adaptive, and creative, and Agent Smith infuses his dialogue with human emotions such as anger and disgust. 

With the line between man and machine blurred to the point of being almost non-existent, the Matrix trilogy raises the complicated question of how interdependent man and machine really are.

3. The Terminator in The Terminator Franchise (1984-2019)

 "I'll be back", "Come with me if you want to live", "hasta la vista, baby". The Terminator has some of the most memorable lines and is renowned as the film that kickstarted then-weightlifter Arnold Schwarzenegger's career; it is also one of the quintessential 80s movies. In addition to the blaring neon lighting and eye-gouging fashion, it is one of the best summations of the decade's anxieties about computers and nuclear war. 

The 80s was the first-time computers became visible in most people's lives. Businesses began replacing paper filing systems with digital ones, typewriters were swapped out for Amigas, and home computers suddenly became affordable. This theme of the intrusion of technology into everyday life is embodied in the plot of The Terminator. The movie is literally about a woman whose normal, unremarkable life in L.A. is interrupted by a computer sending a robot to kill her. 

The film's particular brand of paranoia is called "technophobia," the fear of advanced technology. And who wouldn't be afraid of the Terminator? Unlike most other AI villains, the form of this one came not in a slim figure of the Agents from the Matrix, but the much larger, much more formidable T-800. The purpose of the AI in The Terminator is to kill Sarah Conner and, in the subsequent sequels, her son, John Conner.  

The story tells of a world where humanity has lost control of the very tools it created to make its world a safer, more vibrant place. The machines have taken over and can blend seamlessly into human society, making it difficult to tell the difference between the man and the machine.

2. Replicants in Blade Runner (1982) & Blade Runner 2049 (2017)

 Blade Runner may feature the greatest number of AIs and might ask the most profound questions about what it means to be an AI. The film takes place in an imagined 2019, and though it may have overshot the mark in some of its technical details (we don't have any flying cars just yet), it could not be sharper with respect to the anxieties that define our age. 

Sir Ridley Scott, the director of the masterpiece, imagines a world controlled by a few large corporations that have become enormously profitable through the development of intelligent machines. These humanoid robots, known as "replicants," are effectively enslaved and used for hard labour on distant planets. Despite their enslavement, there is a pervasive fear that they will infiltrate other areas of human life. The film tells the story of Deckard, a so-called "blade runner" charged with hunting down a group of replicants who escaped from an off-world colony. Deckard disdains replicants, but in his pursuit, he unwittingly falls in love with one and confronts the possibility that he might be a replicant himself.  

Blade Runner raises the challenging philosophical question of whether a constructed being can, or should, be considered a person. 

1. Hal in 2001: A Space Odyssey (1968)

 2001 is a renowned classic, often ranked as one of the greatest films ever made and inspired the likes of Christopher Nolan, James Cameron, Danny Boyle, and Steven Spielberg. Although depending on your personal preferences, this Stanley Kubrick film is either a work of complete genius or a pretentious, incoherent mess. There is no dispute that it significantly impacted cinema and is one of the earliest films to depict artificial intelligence. 

Aboard the spaceship Discovery One, only the supercomputer HAL 9000 has been informed of the purpose of the mission. Widely considered to be infallible, HAL makes an error, which it refuses to accept, dismissing the mistake as "human error". In principle, humans are the computer's designers, but, if it is to be believed, could it be the computer itself? Adopting this line of reasoning, the machine gives itself a status that crew members could not imagine – that of a living being. 

To surmise, 2001: A Space Odyssey explores technological innovation, its possibilities, and its perils. Hal presents the problems that can arise when mankind creates machines whose inner workings they cannot fully comprehend. 

It’s just a movie, right?

While we may be a long way from cyborgs travelling through time and taking over the world, AI poses some significant challenges for the human race – one that we all need to be conscious of.

AI can be a tool for good in the world – or it can be a tool for bad. For instance, at Sentient Machines, we use AI to help customers and businesses understand each other and communicate better. We believe that artificial intelligence brings many benefits and opportunities for people, but we must be careful and use these tools correctly.

Written by Minul de Alwis

5 Bad Habits to Avoid to Increase Customer Loyalty (And How AI Can Help)

It’s safe to say we have all developed some bad habits throughout our lives, especially when it comes to work. Whether it’s skim reading through important emails or refusing to take a 10–15-minute work break, we all have them.

When it comes to customer service, it is far too easy to fall into bad habits, which can lead to poor customer experience, negatively impacting your churn rate. 

As with any negative trait, your bad customer service habits can be broken, and with the help of artificial intelligence (AI), it is even easier to do just that.

1. Blaming your customers for miscommunication.  

"The customer is always right" – it's a business mantra as old as time, but it is more relevant now than ever. These days, the business that knows their customers best and caters to their needs more often than not comes out on top. It is far too common; however, during customer service interactions, things can get lost in communication – especially if the customer is emotional or distressed.

Your business can leverage AI by utilising natural language understanding (NLU) which provides real-time analysis of customer service calls. This analysis can give you a better understanding of the conversation between the customer service representative and the customer. AI can offer ways to improve the experience via understanding the customer's level of frustration, the need for escalation and quicker resolution of problems.

With AI from Sentient Machines, you can quickly discover if there is too much confusion in your customer interactions or if customers are constantly looking for clarification. Then you can train your team efficiently to deliver a smoother customer experience, even during the most challenging interactions. 

2. Ignore your customer by replying after a few days or, even worse, not replying at all.

A study conducted by Arise a few years ago found that while around two-thirds of consumers would accept hold times of less than two minutes, 13% said that no hold times at all were acceptable. Moreover, about 34% or one-third of callers hang up and never call back if their call is not answered within a reasonable time. This is also known as Abandonment Rate, one of the more critical call centre metrics.

We have all experienced poor customer service, waiting for hours to resolve simple issues. There is no excuse for letting your customers wait in this day and age, with so much advanced technology at our disposal. And especially with so much competition, there is no room for ignorance if you want to run a successful business.

AI can now let you measure customer wait times. Historically, poor service has been difficult to track. With AI, you can collect actionable insights on each interaction and use that transparency to perfect your customer service.

With the Sentient analytics platform, you can automatically group conversations by topics and use it to draft a reply to your customers within seconds. You could even set alerts to reply to your vulnerable customers quicker than the competition.

3. Charge customers for something that does not generate value.

Charging high fees for doing a lot of work invisible to your customer is a thing of the past. If you need to do a lot of setups that your customers can't see, ensure this is reflected in the price.

With AI, you could gather such early signs in customer interactions based on how they phrase questions or talk about benefits.

4. Sounding incompetent.

This is a tricky one. I believe everyone has their place in the world, but it can sometimes take a few trials and errors before people understand what they want and whether they resonate with the company's vision.

In the meantime, those people might have a few conversations with your customers, so you need to make sure your customers don't get a perception of incompetence. It is better to strengthen your interview process to ensure you are hiring people who are 100% aligned with your vision for the future in all aspects.

This means that it's harder to grow your team quickly, but what you gain long-term is more consistency and competence, and your customers will undoubtedly be able to feel this. With AI, you could help your team understand their strengths and weaknesses and upgrade your training system to make it fully bespoke in a matter of seconds. 

5. Lock in your customer and keep referring to terms and conditions.

Have you ever read terms and conditions? You are not alone, not many people do. A recent survey from Deloitte found that 90% of consumers accept legal terms and conditions without ever reading them.

We are experiencing the end of the terms and conditions era and the beginning of humane customer relationships. If you are referring to your T and Cs in talking to your customer, you've already done something wrong. The new era is about customer loyalty and making sure you amend T and C on the fly if they are not suitable for your customer.

The best utilisation of AI is not to replace human interaction but to enhance it and decrease the friction in the customer experience. With our platform, you can highlight interactions where agents refer to terms and conditions or discuss in too much detail elements that customers don't value, resulting in low customer satisfaction.

It's never too late to break a habit.

Speech analytics software greatly improves contact centre performance when combined with human intervention. Apart from contact centre AI technology, human agents can listen to and transcribe recordings while spotting errors and inconsistencies just as effectively. What is more, they can stamp out these bad habits that may be hindering their customer service.

With AI, you can put those bad habits to bed once and for all, creating a better experience for your customers and, in turn, creating greater customer loyalty. 

Dispelling the Myths about Speech Analytics

Speech analytics platforms capture the pulse of customers, enabling enterprises to extract a wealth of information from every engagement and improve the quality of the customer service they provide. Despite gaining momentum, prevalent myths still remain around speech analytics. 

This article will dispel some of the most common myths that prevent organisations of all sizes from successfully utilising speech analytics and unleashing their full potential.

Myth #1: People speak using perfect grammar.

When writing an email or crafting a letter, we tend to be meticulous. There is a cognitive mechanism that happens during this process as thoughts are summarised and organised. We plan, we write, then scan for errors before sending. However, the same is not true for speech.

Mark Twain once said, "I didn't have time to write you a short letter, so I wrote you a long one." But even analysing a long letter would be easier than analysing speech. Why?

Because people don't articulate their thoughts as perfectly as they usually do in writing - they 'think aloud'. In response to a plain "Yes or no" question, I recall one conversation where a customer said, "No, no no, I mean Yes"—now, getting a computer to fully grasp the reason for "no no no" before yes, will be near impossible. Speech analytics tools are impressive, but not infallible, bringing us to myth number two.

Myth #2: Speech analytics is a set-it-and-forget-it technology.

The fundamental goal of speech analytics is to unlock the treasure trove of insight and data housed in the many conversations with customers pouring through your contact centre each day. Without speech analytics, these valuable insights remain locked behind call recordings as none of us have the capacity to listen to all recorded calls & identify recurring patterns.

That doesn’t mean human analysis is redundant. Quite the opposite, in fact. It’s crucial to understand that the insights surfaced are useless without human action.

Speech analytics is hands-on. The more you use it and understand it, the better results and greater the return.

Myth #3: Speech Analytics is too expensive.

If you have been scared off by the steep price tag for speech analytics, it is time to look again. Speech analytics has never been more affordable!

Advancements in cloud-based solutions have eliminated hardware expenses, while out-of-the-box operations end the need for expensive professional services and implementation fees. The cost of speech to text transcription has dropped, while speed and accuracy have increased, meaning you get more bang for your buck. 

Myth #4: It takes 6-12 months to get to ROI.

While speech analytics has been around for the past few decades, it hasn't garnered much of a reputation. We could blame this on many factors. For one, it largely relies on the accuracy of speech to text, meaning it isn't 100% accurate. Another prevalent myth about speech analytics is that it takes time to see results, with up to 6 to 12 months before you see any return on your investment.

Analytic software can be trained to target key phrases and emotions expressed by the customer, allowing a business to analyse their mood, interaction, and overall satisfaction in real-time.

Speech analytics will save you time and money from countless hours of manual call analysis, help you improve your overall customer interactions and relations, and have a near-direct impact on your ROI.

Myth #5: Speech Analytics is only for the contact centre and large companies.

Let's quickly dispel the myths of exclusivity. While it's true that speech analytics can boost the quality assurance and coaching efforts for any contact centre, this isn't exclusive to customer support. How can other industries benefit from speech analytics?

The travel and hospitality industry - an industry that for the most part has been phone and app-based - can use speech analytics to reduce overheads in booking services and improve customer experience according to tastes, dietary requirements and cultural preferences. The health industry may also begin to see benefits from speech analysis. While not possible just yet, using voice data to mine health-related information about the caller will be viable in the not too distant future to provide appropriate emergency responses to those in need.

Myth #6: Speech Analytics Software is Too Complicated to Use

It is quite natural to feel daunted by the challenge ahead when installing new and unfamiliar software. Legacy speech analytics platforms can be quite challenging to use, requiring you to master the skills of finding interesting insights. Yet, speech analytics software does not have to be that complicated.

The software is extremely powerful and often only needs an Internet browser to run, however there are different interfaces available. At Sentient Machines our goal is to make our UI as easy to use as possible by our clients, which means that they are only a few clicks away from gaining transformative insights. There is no need to have any prior knowledge of how the system works or what Artificial Intelligence is.

Myth #7: Speech analytics is the same as speech to text.

I have often heard people refer to speech analytics as simply transcription, and not fully understanding the difference between the two. Let’s break it down.

You can define speech analytics as a collection of programs and statistical algorithms that help to analyse live or pre-recorded phone calls and gather structured data from unstructured conversations. On the other hand, text analytics is a technology that helps extract meaningful and structured information from written text by equipping the machine to decode and understand a human-written natural language.

Speech analytics will use various voice parameters that reflect changes in the autonomic and somatic nervous system to detect different emotions, moods, and stress levels. For instance, the loudness of voice can be correlated with anger, while fast speech could suggest frustration. On the other hand, text analytics uses words and phrases with positive or negative connotes and disfluency words to gauge emotion.

Conclusion.

While speech analytics has been around for the best part of a decade, businesses are only now starting to understand how valuable speech analytics is in enhancing their strategic planning processes and improving their business strategy.

With this technology, you'll be able to see what's working and what's not for your company through various metrics – such as customer satisfaction rates or customer engagement time on your site.

5 Ways AI Can Help Remove Miscommunication in Call Centres

Language is one of the most intuitive ways humans use to communicate, alongside gestures and visual elements. Yet, we often take it for granted. 

Subtle factors that we generally consider insignificant, including our education, culture, previous experiences, and emotional states, influence how we use language as an effective communication tool.

Misunderstanding others is a leading cause of frustration across all areas of our lives - among our family, friends, other loved ones, or employees, colleagues, and customers. In the case of business, misunderstanding is fraught with danger to reputation and profits.

This blog post explores five elements of AI that you can use to improve your customer interactions and eradicate miscommunication among your team.

1. Understanding nuances of speech.

Psychologist Elizabeth Stokoe has spent the past 20 years studying the science of talk. During her career, she has analysed police interrogations, public mediations and everyday conversations for the pauses, nuances and details that can provide insight into how we can communicate better. 

She explains that how we use language and interact with each another can be used to predict the outcomes of relationships. For instance, pauses - the human ear can't quite hear them, especially when we are engaged in a conversation. Prof. Stokoe uses a brilliant and somewhat amusing example of how the pauses in speech can predict whether a couple will break up. See her TED talk here. 

The human ear has a tremendous range and sensitivity, but it cannot pick up the nuances of speech. AI — specifically Natural Language Processing (NLP) — is advancing, providing machines with language capabilities, opening up a new realm of possibilities for how you can communicate with customers.

2. AI can identify whether "grounding" has been established between two speakers.

Lack of grounding is one of the leading causes of miscommunication. Grounding is a fundamental aspect of spoken language, enabling humans to acquire and use words and sentences in context. In Natural Language Processing (NLP) it refers to establishing the ground truth - a confirmation of whether a speaker understood the other one.

This may mean an agent repeating customer requests in their own words in a call centre, or a customer seeking a confirmation from an agent that they understood what has been discussed.

When communicating with someone over the phone and the two speakers come from different backgrounds and, in some cases, from different parts of the world, finding common ground can be difficult. This is where Natural Language Processing (NLP) comes in.

NLP is a subfield of artificial intelligence (AI). It helps machines process and understand the human language to perform repetitive tasks automatically. Examples include machine translation, summarisation, ticket classification, and spell-check. NLP helps machines make sense of human language faster, more accurately, and more consistently than human agents. While there are still many challenges in natural language processing, its benefits for call centres are enormous.

Being able to establish grounding through the use of NLP, means that you have to deploy a sophisticated dialogue act engine and algorithms capable of understanding not only what is being said, but also who has said what, and when, and what are the implications of this.

You can learn more about basic concepts of NLP here.

3. AI can be used to assess if one party feels being 'heard'.

When conversing over the phone, it can be challenging to establish whether there is a natural flow of conversation. However, AI today can be reasonably accurate in assessing whether one party is listening to the other. This is essential in the call centre environment as the customer needs to feel 'heard', and goes beyond mere measuring if one party speaks over the other.

Our cutting-edge Sentient Analytics platform will measure your agent’s ability to listen and understand, giving them useful clues on where and how they could improve.

In addition, it can notify you what customers like or dislike about your products or services, enabling full transparency to your customer interactions and giving you the power to keep customer needs at the heart of every conversation.

4. AI can gauge if there is an emotional charge to the conversation.

Artificial Intelligence is being used to assist staff in identifying upset clients and de-escalate problematic situations.

Research from Forrester, carried out on behalf of CX software firm CallMiner, found that 67% of call centres surveyed are dealing with more complex customer requests than the previous year, with 70% facing an increase in calls from emotionally charged consumers. 

If there is an emotional charge, speakers will focus more on their emotions rather than finding a solution. 

Our advanced conversational analytics platform, Sentient Analytics, can extract not only positive and negative sentiment, but a very granular emotions from the caller, giving you a better understanding of their feelings.

5. AI can be an 'emotion free' listener giving out objective views.

When we are in a heated conversation, it’s often hard to see a bigger picture. If a third party is listening to the conversation, they can immediately notice what and where went wrong and explain why there was a miscommunication.

AI can act as the silent listener highlighting objectively different elements that impact the flow of conversation. Is it a lack of empathy? Confusion? A concern? Fear? 

The Sentient platform can extract granular elements such as empathy, confusion, anxiety, concern, and more than thirty other emotions, derived from millions of your customer interactions.

Conclusion.

Customer care and support are already among the leading use cases for AI technology, and 73% of industry leaders polled by MIT expect AI's role in this area to grow over the next few years.

Using AI in call centres aims to remove any language barrier, limit the chance of miscommunication, improve the customer experience and relieve human agents of time spent on simple requests. AI can help customer support agents be more productive, have engaging and personally satisfying conversations with their customers.

You can check out this short video from Sentient Machines here.

7 Tips and Strategies for building Award Winning Call Centres - and where speech analytics can help

I was recently a victim of identity fraud. Upon finding out, my initial feeling was one of dread as I realised I would have to call my bank.

After explaining my situation to the call agent, I was routed to the ‘fraud’ team. So far, so good.

“Why did you ring US?”

“No, I didn’t ring, I was transferred by your colleague because I’ve been a victim of identity fraud.”

Then I have to explain the entire situation again…

“Sorry, we can’t deal with that here. I need to transfer you.”

Then the line goes dead.

This encounter and others much like it are the reason I avoid calling them. If a situations arises again in the future and I’m forced to call once more, I’ll probably consider switching banks as that might be less painful.

Sadly, we’ve all been there. Trying to get through to customer service with a seemingly simple issue or request, only to be told after a long wait time that you need to call a different number, or the department you need is now closed, or some other frustrating reason why your request cannot be actioned.

I’ve often wondered what the ratio between wasteful vs. productive time would look like if we (honestly) calculated this for each person engaging in these types of conversations.

Of course, not all companies are the same, and not every call centre is broken. But the majority still are.

So the question is, how do we fix it?

Here are a few tips derived from an insightful discussion (more here) with a true expert who has managed many award-winning call centre teams, Dave D'Arcy. 

Tip 1: Forgiveness is easier than permission.

Tim Ferris in his “The Four Hour Work Week” speaks about the power of giving your team authority to make decisions that don’t cost more than a set budget. For Tim, doing this freed up valuable time to handle more important tasks and gave his team a feeling of empowerment.

Dave goes a step further to remove any set monetary constraints directly, but requests his team simply act as if they were the business owners - hence taking more reasonable decisions and potentially offering a better outcome for the business long term.

The idea being that if you go to your supervisor asking for permission, for one reason or another they might try and talk you out of it. But if the team have authority to solve problems by behaving as if they were the business owner, whilst they may make mistakes you can forgive, more often than not they will find a solution that works well for the customer and the business.

In summary, give your team authority and they will go out of their way to help the customer, taking full responsibility for immediate resolution.

Tip 2: Doing what is right

Most of the time, customer-facing staff are under a lot of pressure to hit targets, while at the same time exceed customer expectations and be compliant. However, often what is right for the customer, is in conflict with company procedures or systems. 

How do you ensure your staff does what is right when there is a possible conflict with the process?

If the process is wrong for what the customer requires, consider that it is probably wrong for the business, and the protocol should be challenged. In allowing something to be handled differently, you will likely improve it. The challenge here is creating the right environment for front line staff to feel confident amending rules for exceptional cases, and then having an easy way to send these requirements back to the business so that the procedures can be updated. 

Today, you can gather very sophisticated insights using AI and conversational analytics to not only understand communication skills and training needs for your call agents, but for this particular case, a sophisticated AI solution could help bridge the gap between the front line and business.

Tip 3: Happy staff means happy customers

Most companies are understandably focussed on improving service in order to increase CSAT and/or NPS. These same companies would likely also agree that happy staff make happy customers.

But if that’s the case, why don’t we start with staff happiness when trying to improve customer service? Their engagement, morale, motivation, mental health - particularly in the wake of a global pandemic.

“If the general mood of my team is dipping, I’ll see a dip in outputs such as CSAT” says Dave D’Arcy.

Dave encourages shifting focus to the inputs, but how do you measure these? By doing what Dave calls “regular temperature checks”. Reaching out to people to say “how are you?”, making them feel human and valued.

You cannot simply tell call agents to be happy or to use positive words and hope this will be enough. As a leader, it is your responsibility to give your agents a reason to smile on the phone.

If you’re able to achieve this, the next step is monitoring it at scale because anything that cannot be measured cannot be improved.

With the right speech analytics solution, you can be very efficient in understanding whether your staff exude genuine happiness through their tone of voice, or whether they are just reading ‘amazing’ words from a script.

Sentient Machines platform uses more than 25 AI algorithms to assess each agent’s ability to listen, show respect, be helpful and friendly to the customer, and focus on solving their problem.

Tip 4: Know what your AI is for and be honest about it

How do you go about choosing technology such as chatbots, voice bots/automated IVR, or conversational analytics? How do you decide which tech to use for your team?

There are many use cases where AI can be useful when it comes to supporting customer care teams, but as we know, it can also be overkill and lead to customer frustration if they can't speak with a human when needed.

While changing the input is going to impact the output, it’s always important to know the outcomes and measure them thoroughly. To convince management that staff happiness is the way to go, you will undoubtedly need to prove that the outcomes are in your favour. Is your business gradually becoming more profitable with AI? Are your CX measures going up? NPS, CSAT, CES?

When implementing any new technology, it is always important to ask “What do you want it for?”. Is Customer Effort going to be less with AI? Great. If the requirement is to improve CSAT, make that the focus. Similarly, if its only purpose is to reduce costs, that’s also fine, so long as you are clear about it from the start and don’t expect the AI to solve a problem you never intended for it to solve when you adopted it.

Tip 5: Work smarter, rather than harder

For a long time, call centres have been measuring their team’s performance based on very dry metrics. The number of calls, the handling time, number of calls per day. We can all agree these are two-dimensional and ineffective if we want to improve the performance of call agents.

With speech analytics, you are able to go very granular on a large scale in understanding the impact of your agents’ performance such as soft skills, upselling, and much more, rather than simply counting the number of calls they took.

It’s about working smarter rather than harder. And gathering insights from a sophisticated speech analytics platform can be priceless in this context.

Tip 6: Understand your team’s desires

Covid taught us many lessons. Above all: How to facilitate working from home effectively. 

In many call centres pre-pandemic, working from home wasn’t even an option. The pandemic has forced everyone to revisit that, and companies have become more flexible and more open to this idea.

Perhaps again the way to start is by asking your staff what their desires are. If working from home helps an individual with their well-being, then finding a way to accommodate that will have a far-reaching positive impact on the agent, their performance, and ultimately your customers.

Tip 7: Bridge the gap between the front line and back end team

Engineers and R&D teams are often not very excited about fixing customer bugs, as they’d rather work on more exciting new features. However, the only way to improve CX is to create an alignment between the two.

I’ve often seen the two departments completely separated and it always seems a waste of resources not to connect the dots.

While analysing customer interactions with our advanced speech analytics platform, we’ve discovered topics that emerged directly from a customer’s experiment of using a product, and they call or write about it in such a detail that it is a missed opportunity not to be able to close this loop. Some customers even call solely to offer feedback on a product and service.

It reminds me of when, while being pregnant, I managed to get Covid. So I was keen to share this with the hospital managing my pregnancy because I thought it would be a really valuable piece of information, as at the time there was little evidence of how Covid affected pregnancy. They just stared at me before saying: “I don’t have anywhere to input that information”. 

And it’s the same with call centres: they’ve nowhere to input the information that can go to relevant teams and help improve products and services.

But if you have speech analytics analysing every customer interaction, the insights can go directly to their inbox. It’s as simple as that.

The Truth about Innovation in AI

In February 2020, I read this article describing the initial vision of OpenAI and how it had evolved. I find the whole concept around OpenAI quite exciting, and especially the vision which is focused on the ethical AI, and how to make sure it does not go all wrong at the end once AGI (Artificial General Intelligence) arrives. AGI arriving means machines have the abilities to fully simulate a human mind, which has brought many questions and worries to date. Hence, having an organisation that is on the mission to make sure ‘that the technology is developed safely and its benefits distributed evenly to the world” sounds good to me.

But the article was quite bold about innovation — or the lack of it — within OpenAI:

“Most of [OpenAI] breakthroughs have been the product of sinking dramatically greater computational resources into technical innovations developed in other labs.”

In other words, this article was explaining that there has been a lot of investment into OpenAI, and all they’ve done was paid for servers on which to run someone else’s ‘innovation’ so they can get some results that they’ll then want to call a breakthrough.

Yet, a few months later, OpenAI released what is considered one of the biggest breakthroughs in AI in decades called GPT-3.

How is this possible?

What is wrong with ‘re-cycling algorithms’ in science?

Judgement around OpenAI’s innovation came from the fact that they were reusing what has already been built in other labs, despite the fact that they were by definition working on pushing the AI game forward. The question is: What exactly is wrong with reusing breakthroughs and building up on them?

Over the years, there are some libraries that I have been so grateful to be able to use, because they’ve made me 90%+ more productive. Sometimes you have to do things from scratch, but there is a reward from re-using as well as learning from someone else’s work and experience. Someone who has possibly have done it from scratch and had to figure out all those little things that always come to be on your way, and you have to keep going, keep thinking, keep overcoming them to move to the creative side of whatever you are trying to accomplish.

Can ‘recycling’ science then be considered innovation?

And would it have been wiser if OpenAI ignored everything that has happened in the world and started from scratch?

Any type of recycling is always a positive, and even more so if this enables you to move things forward. If you are enabling others to make progress. If people can learn from your ‘recycling’ to enable further innovation. As you are not necessarily recycling. You are building up on top of someone else’s work.

Sometimes, innovation requires deleting everything you know and starting from scratch. Other times, innovation means building up on what is already out there. Which is, I’d assume, why OpenAI was (and I’m hoping still is) “sinking dramatically greater computational resources into technical innovations developed in other labs”.

In conversation with Dr Xingyi Song, a Speech and NLP scientist, he mentioned that “recycling an algorithm” in AI is similar to a physicist doing experiment to prove an unproven theory, and that models such as GPT-3 enabled us to know what the deep learning, or transformer models are capable of.

Prior to GPT-3, BERT model released by Google was a game-changer for anyone working in Natural Language Processing. It is estimated that the cost for Google to train one iteration of BERT model is up to $1.6m (see this paper for more details). Practically speaking you’d probably need a few iterations, which means it might cost you around $10m to use BERT in your application if you were to train it from scratch. Who else has that much money to spare? And who else has access to all the data that Google has at hand?

What innovation isn’t?

Reinventing the wheel. Building things from scratch if you don’t have to.

To teach me how to properly serve in tennis, my coach said he first had to make me forget what I knew. This left me having two ‘left’ hands for a few weeks, but then at least my serve started to look like I’m not playing badminton. Starting from a tabula rasa is often useful, especially when you’ve realised that you won’t go very far unless you change something, no matter how much resources you put into it. But, if my serve was decent to start with, perhaps investing more time and hours and practice would get me a breakthrough?

When you have to start from scratch

Starting from scratch is what Musk had to do with Tesla. He made a car that isn’t really a car, in a traditional sense. It’s a super-computer. It feels alive. Other car manufacturers tried to make electric cars by building up on the knowledge they already had from building millions of their petrol and diesel cars. But their knowledge wasn’t helping them, because they’ve spent too much time in the car industry and it was difficult to start from scratch for a project that needed a full reset. An electric car, and a self-driving car, was craving for a new innovative way to completely disrupt the car industry. It needed to be done from scratch. Musk made it so cool that even someone who doesn’t know anything about cars (me) can get really excited about Tesla, as it is obviously the future. As a technologist, I can see that there is no turning back. Musk has already made a dent, now it’s only a matter of how quickly can it be made affordable to everyone.

Even with Musk and Tesla, I would argue that, while it can be seen as a completely “from scratch” disruptive project, all those hours and days and sleepless nights are the ones that Musk had to invest in order to build the foundations for his learning (see this article to appreciate why this worked). To start from scratch, you really need to have your foundations steady.

Algorithms are only part of the puzzle

We didn’t really see a massive change in the algorithms since the beginning of AI in 1950s, but what we have seen is the amount of data we suddenly have available, as well as the computational resources that we can use to process the data through. Most AI breakthroughs were a result of building up on the knowledge we already knew, but figuring out how to do it on a large scale (take for example IBM’s breakthrough with Watson’s winning a Jeopardy competition outperforming humans in 2011).

If some innovations have already been generating positive results, there is nothing wrong with trying to push them even further, when you are trying to make a breakthrough.

Plus, no technical innovations developed in other labs (assuming this refers to open source libraries such as Tensorflow) are capable of solving the problem by themselves, without putting significant resources into engineering and use of the actual algorithm and also, cleaning and preparing your dataset, plus deciding on what your dataset is, which part of it to use, and how. And not to forget, defining what you want your algorithm to do. All of these are massive decisions that need to be made in order to do anything with “technical innovations developed in other labs”.

Speaking of Tensorflow, I remember Google’s announcement of it going open-source and people asking them “Why did you do this?”. They said: “Because we can’t do it alone”. While it is arguable that there are many other reasons that could apply, it is 100% correct that even Google, with unlimited resources comparatively to anyone else, can’t do it alone.

Given the success of GPT-3 released recently by OpenAI we could argue that all we need is to “[sink] dramatically greater computation resources into technical innovations developed in other labs”?

If only that was so easy.

How to Craft Emotion-centric Customer Interactions? — Part 3/3

How to Communicate Effectively with Your Customers

In the previous post, we discussed how What you do impacts the quality of your customer interactions and your bottom line, and how analysing your conversations in real-time can help transform your customer satisfaction. Today, we are focusing on the What you say — the communication aspect of your customer interactions.

We analysed over 900,000 customer interactions, and derived five elements of communication that made an impact with regards to successful sales outcomes and higher customer satisfaction:

1. Empathy

Empathy is a very popular one, and there is a lot of advice online and in literature on how you should use it in order to achieve high NPS and customer satisfaction scores. In the first post of this series, we’ve seen however that significant empathy doesn’t help with removing low CSAT resolution scores. Which brings back the importance of delivering quality service and products to your customers, before you do anything else. In other words, empathy is not a magic ingredient, and if your products and services are not satisfying for the customers, improving them should be your number one priority.

5 elements of effective communication in contact centres

We analysed inbound support calls with the Sentient Machines platform and extracted those with emphatic agent behaviour, then correlated these with Customer Satisfaction (CSAT) Agent and Resolution scores from our customer.

• CSAT Agent — reflecting whether the customer was satisfied with the quality of the interaction with an agent,

• CSAT Resolution — reflecting whether the customer was satisfied with the resolution.

For those interactions that received the highest CSAT Agent score of 5 (see Figure 1), there was as much as 33.9% of empathy, which can lead us to a conclusion that empathic agents have a high chance of receiving a very high CSAT score.

For those interactions that received the lowest CSAT Agent score of 1, there was only 0.66% of agent empathy. However, we see that almost 24.5% of empathy is found in calls with the lowest CSAT resolution score of 1.

Similar to what we’ve discussed above, Figure 1 highlights the importance of resolution in addition to training agents.

Figure 1: Absence of Agent Empathy Drives low CSAT Agent Scores, while significant Empathy Doesn’t Help to Avoid Low CSAT Resolution Scores

In Figure 2, we demonstrate results of agent empathy in sales calls. More empathy is present in successful sales — 15% as opposed to 8 %, and this is inline with most studies you will find in literature.

Figure 2: Empathetic language drives sales

However, in another study (Figure 3), we observe the opposite impact. In this instance, empathy is present almost 2% more in unsuccessful sales compared to successful sales calls. When we analysed why, we found that these conversations included retention calls — conversations with customers who were enquiring to cancel, while the company tried to retain them by offering better deals.

Figure 3: More empathy in unsuccessful sales due to agents increased effort to retain customers.

In Figure 4, we plot different emotion algorithms extracted with the Sentient Machines platform, against agents and their sales outcomes. The red line depicts the sales success rate per agent, and the blue bars demonstrate the presence of a particular emotion. We can quickly observe the exact same trend of the red line reaching the peak and then going down, with some arriving at the peak faster than the others. The lesson is that we need emotion, but you don’t want to be dragged into it. To be more effective, you need to acknowledge what’s happened and move on, proposing a solution.

If you are validating your own mistakes for too long it might become counter-productive.

Figure 4: Trend of emotion in calls driving sales outcomes

In conclusion, empathy is great, but not enough. If over-used, empathy can be counter-productive.

2. Personalisation

The next important aspect of effective communication we derived is Personalisation. In Figure 5 we present different techniques we measured for 9 different agents, whose sales success rates range from 6.63% to 33.62%.

What stands out is that Agent 1 was the only one who applied Personalisation technique. They knew everything about customer’s buying behaviours, what the customer liked or disliked. This knowledge resulted in Agent 1 having 3 times more successful sales than the average of everyone else, and over 5 times more successful than the least successful agent, which is outstanding.

The absence of emotion, and instead showing that you care through knowing your customer preferences, has a potential to create even deeper connection.

Figure 5: To Be Emotion Centric, Learn about Your Customer History

3. Reassurance

The next important communication element is Reassurance — Reassuring your customers early in the conversation that you are:

• Focused on solving their problem.

• Focused on their benefits.

• Removing any risks.

At the same time, you need to be actively listening.

4. Upbeat and Confident Tone of Voice

From analysing the data, we found that positive language and upbeat tone is more effective than positive language alone. It is also important to show active customer engagement e.g. by asking them questions.

Fake is not cool

Agents who tried too hard to follow the ‘positive’ script, didn’t come across genuine and this impacted their final success score. Those who acknowledged the company failures and moved on, on average achieved better results.

5. Limit Small Talk

While there is a mixed advice on this one in literature, we found that while you are trying to limit the small talk, you can still be friendly, as too much of it can irritate your customers who could be busy. You can talk about their holidays if they are driving that conversation, but the key thing to remember is:

• Listen to your customer, acknowledge their views and emotions, and move on. Your focus should be quick resolution.

In Summary

Keep listening to your customers and improving your product or service based on what your customers say. Train your team to be exceptional in their communication with customers so they can represent your company culture well:

• Use empathy in genuine ways, and avoid using scripts.

• Know everything you can about your customer history and preferences, and show your customers that you genuinely care.

• Reassure your customers early in the conversation about their benefits and removed risks (e.g. “you can cancel any time”).

• Use an upbeat and confident tone of voice that is genuine.

• And finally, use small talk with caution. Your customers might be busy.

If you want to trial Sentient Machines platform on your data, register your interest for our end of summer offer here.

How to Craft Emotion-centric Customer Interactions? — Part 2/3

On the quest to becoming remarkable in your customer interactions

In the previous post, we discussed the importance of crafting emotion-centric customer experience, and we presented how positive language and Sentiment Index impact the overall customer satisfaction, and sales outcomes. We’ve also seen that no matter how positive and skilled your agents are, your customers want quick resolution in order to give you high customer satisfaction scores (CSAT). One of the key learnings was that to bring your customer satisfaction up you need to eliminate the negative emotion in the first place.

A natural next question is: How do you eliminate negative emotion? In the next two posts, we’ll dig deeper into this, on the way to crafting an emotion-centric customer experience, first looking into Actions — what you do, and secondly looking into Communication — how to communicate with your customers.

What You Do — The Importance of Being Remarkable

In Oversubscribed, Daniel Priestly talks about the importance of being Remarkable. Remarkable with your product, with your services, with your customer interactions. I’d extend this further and encourage this concept outside of your business into your private life. If you strive to be remarkable with every step (or breath) you take, this will then establish that habit of striving for excellence that will soon be embedded in everything you do, and naturally become visible in your product or service.

What this means is that step zero in having exceptional customer service is to eliminate the interaction by giving your customers an excellent product or service. And if you can’t fully eliminate the interaction due to the nature of your business, at least you can be sure that negative interaction is minimised.

How do you maintain that quality in the ever-changing market with new products or features always being added or updated? How do you know your customer preference and what your customers like or dislike?

What You Do — The Importance of Understanding your Customers

A lot of companies collect customer feedback post-interaction in order to revise and improve. While post-interaction feedback is better than no feedback, it lacks the real-time element, and by the time you see the causes of customer frustration, your customers might have already churned. While post-interaction customer feedback could complement real-time interaction insights, if you are interacting on a daily basis with your customers, these interactions could be a gold mine for your bottom line.

In Figure 7 we see an example from Sentient Machines, grouping negative calls by topics.

Figure 7: Hear Your Customers: Sentient Machines real-AI platform for discovering insights from your customer interactions

We can see that ‘App’ is causing most frustration for customers, but what about the app? Drill into the bubble view to learn that the app doesn’t work on Huawei phone, and that customers have trouble downloading it, and communicate this to your development team in order to eliminate this negative emotion — to eliminate this unnecessary customer interaction at the first place.

Your app should simply work, and if it doesn’t, your priority should be to fix it if this is what your customers need.

What You Do — The Importance of Reacting to Your Customer Needs in Real-time

The quality of your product and services is critical to the success of your business, and learning from your customers in real-time about how you could improve could be transformative for your bottom line. Automated analysis of your customer interactions can help you answer questions such as:

• What business needs to do today to eliminate root causes of customer frustration?

• Why are customers not buying?

• Which features excite customers and which don’t?

We’ve already touched upon using advanced AI conversational analytics such as Sentient Machines to extract root causes of customer frustration. In a similar way, you could adjust the platform to look for answers to other questions. For example, if during your sales calls your agents ask customers for reasons for their cancellations, you could have these grouped, and quantified, so that your executive team can make more data-driven decisions and use them to gain competitive advantage.

With real-time aspect of analysis, you would be learning every hour about customer preferences, without having to wait for your post-interaction customer feedback analysis.

We’ve seen up to 60% churn reduction when analysing customer data, simply by alerting the customer of highly negative cancellation calls in real-time. In the case where the Head of Contact Centre Operations reacted to these within 2 hours, 49% of customers were brought back and eventually didn’t cancel. Now, imagine if they reacted a week or a month later? Those customers would have been already lost to a competitor. Timely reacting to customer interactions insights is key and can be transformative to your bottom line.

While confidently making improvements to your products and services, you should at the same time take care that your team is well trained to understand the company vision and make sure that every customer feels valued when they interact with you. This communication element is a big part of the company culture and could be transformative for the success of your business.

In the next post, we’ll present 5 key points extracted from a case study where we analysed thousands of conversations and looked into what worked and what didn’t with regards to Agent Communication in order to improve sales outcomes.

Covid-19: 5 Tips for Startup Success in Challenging Times

Embracing the new normal to transfigure your trajectory

Covid-19 is one of the biggest challenges in our lifetime and we need a paradigm shift to thrive. I recently gave a talk at an exciting London fintech startup, on redefining challenges to drive success. I predicted that way we communicate and adapt our business in defining moments determines the eventual winners. Our ability to accommodate a new normal means being super-agile; let’s not pretend that nothing has happened. What can a start-up founder do right now to turn this challenge into transformative change?

In the early days of my career, both completing my PhD, and subsequently in industry, I managed projects and people in situations I’ve never experienced before. I had to adapt. The team needed to be motivated as external stakeholders rightly demanded strong results. Super Agile Focused Communication was key, with a core passion to drive success.

Tip 1 — Keep the Focus

I’m slightly obsessed with focused productivity. As an AI Scientist I develop solutions for industry, transforming cutting edge AI into actionable insights. Hence, when I’m solving a particular problem, I’m focused on the unknown unknowns. The upside is getting a lot done without distraction. The downside is demanding too much productivity from others and potentially hurting their feelings. Which I did. Once. Never after that. And this was, simply because of the way my words were interpreted and perceived. It was not because of my intention. My intention was to get the job done as quickly as possible, and with as little distraction as possible. Keeping your focus on the solution, and communicating the changes in a way that the whole team is on board will make you unstoppable.

Solution: Working from Home, with the ‘Rope Trick’

Empathy is a little used word in business. But it’s key to working with colleagues and demanding clients. Aligning team focus, emotions, and big picture dreams, with new goals brings a powerful synergy. Your responsibility as a leader is to share your vision, and encourage others to contribute to the new solution.

“Rope Trick”: Seek to understand the other person’s perspective and act as one.

One of my favourite books — The Diamond Cutter, talks about the concept of the Rope Trick, where you are bound to act as one by a symbolic rope. This means understanding unique personal and business goals, and different personality types. Synergy between teammates necessitates good communication and the goal of mutual success.

After all, we each bring unique qualities to the table and when aligned the sum is greater than the individual parts.

For this to happen, in working from home conditions, communication channels need to be crystal clear. Establish regular catchups between all members of your team, and be warned that some people might take longer than the others to adjust.

You could apply the Rope Trick to the whole team, the whole nation, and the whole world. Which brings me to the second potential hurdle:

Tip 2— The Illusion of Communication: a New Understanding.

“The single biggest problem in communication is the illusion that it has taken place. “

— George Bernard Shaw

In 2011, I worked in a startup building on the technology at the core of Siri before it was sold to Apple. Our goal was to create a learning companion. It was a very exciting project for all of us, and we were so passionate that we went way beyond just improving a core platform.

We made it scalable, without any need to update the algorithms. I remember a day full of meetings with the SRI team in Menlo Park. I later proudly discovered that my UK team had delivered a new ‘learning bot’ on a completely new subject, the same day. SRI team were impressed because they understood the tech. But our non-tech management team didn’t get it. How could they not appreciate what we’ve built? There was a communication issue.

It was two years later, that Google acquired API.AI today known as DialogFlow, and Facebook acquired Wit.ai. Both are now very popular and enable anyone to create a simple bot in a matter of hours. That proved that what we had at that moment was very special, very advanced of the competition, and very hot on the market, but the broken communication lines between the tech and non-tech teams meant this strategic advantage was squandered. The assumption that what we knew, they also knew, was not based on reality. Which is why, to prevent this happening, I decided to lead the business side of Sentient Machines as a CEO.

Solution: Team Communication

Being transparent and as open as possible with everyone is key. In the absence of face-to-face meetings, a video conference can be as productive. Everyone has a part to play in delivering a total resurgence. Acknowledge this, and you will witness a productive environment where everyone is contributing their absolute best to deliver larger goals. Which takes me to the next point:

Tip 3— It’s Your Responsibility to Sell

What are you doing to save costs and make sure you can survive in the next 12 or 18 months? How are you motivating your team and driving sales? The economy will crash significantly harder if all businesses simply stop, and it is our responsibility not only to sell for our businesses, but to support others and enable them to thrive under new market conditions.

Covid-19 has made people vulnerable, and rightly cautious. Your pitch might need to be adjusted. Understanding changing requirements and implementing transformative solutions is paramount.

Solution — Thinking Ahead

As a founder, you always need to think ahead of the game and lead by example. It is important to own your desires, beliefs, and decisions. To always have a plan B for survival in critical conditions.

Ayn Rand in The Fountainhead talks beautifully about Creators vs. Second handers. Creators are typically not welcomed by the world because they disrupt Second handers. But the world is making progress on the back of the work of the Creators. For the benefit of humanity, Creators are what moves us forward.

In the time of crisis, Creators will be the ones making a shift. There is nothing wrong with being authentic and different. Which brings me to the next point:

Tip 4— Understand Your Motivation

We spend a lot of time running around, ticking off our todo list. We are constantly bombarded with information our brain needs to process. In the time of pandemic, you could get exhausted if you don’t have strong motivation. What really gets you out of bed in the morning?

Solution: Find Your Mountain Garden

“Everyone thinks of changing the world, but no one thinks of changing himself.”

– Leo Tolstoy

Recall when you felt content and happy and align with it. Keep reviewing this until you get to a clear understanding of your motivation.

Yoga, swimming, running or any activity that helps you get out of your mind, will afford you a higher level perspective.

It will help you find your Mountain Garden, — a concept from a beautiful fable about self-actualisation. Following your intuition to align with everything you love. Your Mountain Garden can be hidden at first, but becomes clearer as you tune into yourself, rather than the outside world. I remember reading this book at a very stressful time of my career, and it really helped me to realign with my core values and happiness. As if someone has tapped me on a shoulder and said ‘everything is fine, you are doing fine’. It’s exactly what’s needed in the time of uncertainty.

This ability to concentrate brings clarity to the things that matter, which is my next point, taking direct action:

Tip 5— Keep It Fresh

We need to be inspired to move out of our comfort zone on a regular basis. We are creatures of habit, and by adopting the positive aspects, we can optimise personal evolution.

It’s not easy. Even under normal circumstances, being a founder brings a lot of joy, but also stress.

Solution: Surround Yourself with People and Things That Inspire You

It’s important to always make new friends and take note of people who inspire you. Spend as much time with them as possible. People who have their interests aligned with yours (remember The Rope Trick?). Working in a company where people inspire you is important, because that’s what’s going to push you forward.

Elon Musk is a big inspiration. For his tenacity, big vision, persistence, and the fact that money is the resource to get things done. I couldn’t remove the smile from my face when I heard a song he recently produced. In the words of Elon Musk, if you feel inspired to make a difference:

“Don’t doubt your vibe, because it’s true

Don’t doubt your vibe because it’s you”

Is Artificial Intelligence the Biggest Threat to Humanity?

I grew up on the Isle of Wight and regularly spent my childhood summers at the beach. Between building sandcastles and eating too much ice cream, I would happily run in to the sea knowing the only thing I had to worry about was not getting caught in the strong current going in and out the harbour.

When I was 9 however, that all changed.

Suddenly my biggest fear was not being pulled out to sea and out of my depth, no, my biggest fear became being eaten by a giant shark. Well, any shark for that matter.

Since watching Jaws at 9 years old, I have had a lifelong fear of being in the sea. Now don’t get me wrong, if I ever make it to one of those white sandy beaches in the Maldives that I dream about then I will happily splash around in a few feet of crystal clear blue waters. But the murky and potentially monster-rife seas such as those off the Isle of Wight? No thank you.

My local beach. The circling sharks are just out of shot…

Watching Jaws all those years ago forever changed my relationship with the sea. I know it’s irrational (yeah but is it?), but that’s the power of film.

In preparation for writing this blog, I asked on Twitter if anyone felt they had developed an association with a person, place or thing due its portrayal in a fictional film. One follower responded that she will go to any lengths to avoid getting in a lift alone due to the scene in Silence of the Lambs when a body falls from the ceiling.

Hello, Clarice…The lift scene from Silence of the Lambs

The moment I initially started questioning how film may have impacted public opinion on Artificial Intelligence is when a friend asked if the work I do at Sentient Machines will eventually lead to the evolution of a machine such as V.I.K.I. in 2004’s iRobot that seeks to wipe out humankind.

Whilst the question seems funny to me now, I can’t deny that a few years ago before working at Sentient Machines I may have posed a similar question. After all, as an avid filmgoer I have seen the ‘AI becomes sentient and tries to take over the world’ narrative played out many, many times. The Matrix, The Terminator, 2001: A Space Odyssey, and more recently Avengers: Age of Ultron, to name a few.

The creators of these films imagine a world where humans have lost control of the technology they developed and must fight for survival of the human species. It could also be suggested the writers and directors of these films are predicting a world where these things happen. After all, many notable figures in the world of tech such as Bill Gates, Elon Musk, and even Stephen Hawking have all made warnings against the potential consequences of AI.

But how accurate is the silver screen’s depiction and are the fears of my friend based off of these films warranted? Is AI going rogue, building a robot army and attempting to eradicate humans as likely as finding a dead body on top of a lift or being attacked by a giant shark off the Isle of Wight? Considering the capabilities required to orchestrate a world of machine dominance as seen in these blockbuster films, technology remains eons away from creating anything so multifaceted (and let's not forget, the machine would first have to become "conscious" and choose to rebel).

Yes, there are without a doubt many specific tasks where a well-trained machine can perform at greater speed and with more accuracy than a human, but only because it has been taught by a human how to excel in that particular task. Could you train a robot to kill? Of course you could, but you could train a human to kill also, so really the discussion comes down to ethics*. How we define and monitor ethics within the development and use of AI is something we must never ignore or take lightly, but we should also not let fear impede our technological advancement.

*It's also worth noting that a robot trained to kill is very different from a robot who - of its own free - will decides to kill.

I believe that by harnessing AI’s incredible ability to simplify and accelerate these often mundane or humanly unscalable tasks, we as human beings can thrive. If you look hard enough, stories of AI being used for good are everywhere, unfortunately they just don’t always make for exciting movies (aside from Wall-E of course, the story of a robot tasked with clearing rubbish from the streets of Earth, a personal favourite of mine).

This Forbes list ‘8 Powerful Examples of AI for Good’ details just a few of the many positive uses of AI within healthcare, education, human rights, and the fight against climate change. These examples include drones delivering crucial medical supplies or helping clear plastic from the ocean, as well as facial recognition software used to combat human trafficking.

Undoubtedly, our day to day interaction with some form of AI is only going to increase and it is therefore important we look beyond film and encourage others do the same in order to gain a true understanding of the technology facing us - especially if we want to make use of its full potential.

As for me and the sea...if someone could create an AI-powered device (waterproof, obviously) that alerts me to the proximity of sharks relative to my location, that might just tempt me back in.

Why real-time AI is the key to contact centre success

With contact centre agents at the forefront of a business’ interaction with its customers, it’s no wonder companies spend an initial two to six weeks training each agent before they even interact with a customer.

But once the agent is fully operational, how do we define and measure the success of their performance? By the number of sales made? By whether a customer demands to speak with a superior? By inconsistent and frequently incorrect CSAT scores?

With the evolution of AI-powered speech analytics that transcribe and analyse 100% of calls in real-time, companies adopting the technology have been handed the key to a treasure chest full of invaluable and actionable sentiment-based insights.

Sentient Machines recently spoke with Phil Bruce, a former contact centre coach who, when using our leading edge analytics platform to optimise the onboarding and development of his contact centre agents, saw a 30% increase in call quality.

In the above video, Phil mentions listening skills as key to making a successful sale. Below, he elaborates further, suggesting that a sales agent should first understand the customer’s needs before attempting to offer a solution.

“Not only do our words matter, but also the tone which we use has a huge impact.” - Dr Hyder Zahad (HuffPost)

As well as automatically measuring the listening skills of each agent, the Sentient Analytics platform reads and tracks emotions throughout customer interactions by analysing both lexical and acoustic sentiment - so what is being said, but more importantly how it is being said.

For some, there is scepticism surrounding the use of artificial intelligence within contact centres and whether its advancement could render human beings obsolete. At Sentient Machines, we believe in an ethical use of the technology in order to enhance current processes, not replace them.

By combining AI-mined insights with human analysis, you have the power to create a sophisticated contact centre QA process and drive transformative outcomes.

Interestingly, an Opus Research Survey revealed that 72% of companies believe speech analytics can lead to improved customer experience, 68% regard it as a cost saving mechanism, and 52% trust that speech analytics deployment can lead to revenue enhancement.

Of course, cost must always be factored in, but research by DMG Consulting indicates that speech analytics in contact centres pays for itself in less than one year, and Techtarget reports that it can pay for itself in as little as three months.

“Speech analytics can make substantial, quantifiable contributions to contact centers, but this is only one of the many use cases for this highly flexible and effective application.” - Donna Fluss (DMG Consulting)