Ep # 172: Beyond Conversation Intelligence: Adding the Emotional Layer GTM Teams Need to Win ft. Chandra DeKeyser (Co-Founder, Voicera)

Chandra DeKeyser 00:04 If you have a meeting and you have the transcript of the meeting, you think you know what happened because you have the transcript, but you don't. You miss the emotional dimension. Is this customer truthful about their intention to renew? Is it an 80% close rate or a 10% close rate? Intro 00:22 Welcome to Across the Funnel, where we dig into concrete Go-To-Market moves across sales, customer success, and account management so you can build revenue that lasts. Brought to you by Hyperengage and Dextego. Adil Saleh 00:37 Hey, greetings everybody. This is Adil from Hyperengage podcast again. We're doing a lot of these at the recent times. I know that conversational AI has been changing in a good direction, a very fast pace. Generative AI, a lot of these tooling, they are products that we recently had in the past. They are more towards building a conversational AI layer on top of it, having specialized contextual information to make it work for different use cases, different customer segments, and they're building product out of it, and they're growing fast. Today, we are going to be talking about Voicera. Voicera is a platform of different nature that actually does the same, but it is at a level of different personalization, understanding the patterns, understanding the psychological behavior, exceptional behavior of people, having some sort of a neuroscience layer on top of it. I remember we had this product name. There was a product that was more for HR three years back. We spoke on our podcast, and they were doing a similar thing. So, we're going to be in-depth with the co-founder of Voicera, Chandra, today, and to see how they're helping coaches, how they're helping executive leaders to better empower and enable their team by understanding them a whole lot better at a different level. So, thank you very much, Chandra, for taking the time. Chandra DeKeyser 02:14 My pleasure, Adil. Thanks for having me. Adil Saleh 02:17 Likewise. Chandra, I know that you've done your master's and bachelor's really back in the years, even when I was not even born. I was not even in discussion of getting born amongst my parents. Tell us through how academics has changed today, that has changed in terms of technology, how it has impacted the industry, how is it more connected to the industry. Could you give us a little bit of background, your origin? Chandra DeKeyser 02:51 What do you mean? My own trajectory, you mean? Adil Saleh 02:53 Yeah, your origin, how you started education, from where you've done master's in the late 70s and bachelor's in the late 70s, and then how you got to Voicera. Chandra DeKeyser 03:06 You're dating me, in the sense of carbon dating. So, look, I'm an engineer from Belgium. My parents fell in love with India, so Chandra, but I'm Belgian and I did engineering. Then I did a master's in computer science in Italy. I got a scholarship to go there instead of going to America, where I had to pay. So I wanted to go to the States, which was very expensive. In Italy, they paid me to come and do a master's. Basically, I did a master's in computer science, and I did computer science as a geek initially, writing code in assembly and all that. And I moved towards project management and then evolved into business development, sales for large companies over the years. End of the 90s, I was doing mainly entrepreneurship or directing business units in business dev in Europe for big companies like Northrop Grumman, but totally civilian stuff, government IT systems. Worked for Software AG also for four years, also as head of director of business development. Went back to my first love. My first job as a geek was a startup and the second job was also a startup. And I loved that five guys and girls take decisions and we can change the world if we are very lucky and very hardworking. So I went back to entrepreneurship and I built initially another company called MoodMe, which was focused on augmented reality face filters. And further along, I went into emotion detection. Voicera is more recent. I started that three years ago. I decided to start a company in the US, not in Europe, because I felt that the ecosystem there is more dynamic. All my co-founders, except one, are from the US. We felt that the whole world was going to video, and the video means transcription. And then you have the text and the LLM are crunching the text. And so you know what people said and you can summarize it, action items and all that. So you know what, but you don't know how. And if I tell you, oh, Adil, it's very nice to meet you at the first date, you know that it's not going well. Or as a customer, you say, oh, I like your product, facial expressions, tone of voice. It's universal. You come from all over the world. You go with a shit face. People know you're not truthful. So this is how people say things. As humans, we understand it more or less. Sometimes you are more empathic, sometimes less. But sometimes it is more important. If you have a meeting and you have the transcript of the meeting, you think you know what happened because you have the transcript. But you don't. You miss the emotional dimension. Is this customer truthful about their intention to renew? Is it an 80% close rate or a 10% close rate? By the transcript, it would be 80%. But you're totally going against the wall. It's 10%. They don't like you or you or your product. So this emotional dimension, we felt it was exciting. We felt also it was unique. At that point, everybody was on the massification of the LLM, but not too many people. A few people are in the neuroscience, but not too many. And I teamed up with a guy who was the chief data scientist of SAP, Vipul Patel, very solid Chief AI back then. And so we built a foundation of AI covering sincerity detection, which is our focus, but also emotion detection and also large language model. We've been dabbling with several Llama of Meta. But the focus of the company is the insights on the sincerity of people in video, also audio, but mainly video. So that's where we are today. Adil Saleh 07:01 It's basically capturing the right intent, capturing the right, I would not say psychology, because that's so hard to get into the brains of people and be mentalist and use technology and make a technology that smart. How challenging was that? I know that LLMs and all of these can understand the patterns that have been recorded by humans. But of course, when it comes to unique personalities, every human is different. They think different, they talk different, in a good or bad way. So what is the neuroscience component of it, which you told me at the backstage? Chandra DeKeyser 07:41 So, look. So it's deep learning. So the big challenge is to find a reliable data set. And there are no open data sets, or there are some for research purpose, but we were not able to take an existing data set and train a model. We couldn't do that. So the big challenge was really the foundation to build a data set of people talking with or without sincerity. So that's been a challenge. And it took us 18 months to get to a point where we have a product that is now live. You can open an account on our website. You go to voicera.io and you sign up for free. You have access to our platform. You can send, upload, and get five video files analyzed per month for free. And then if you want to go commercial, you can pay, of course. The big challenge was really building the data set with the right methodology, finding. We decided also to approach the understanding of people in a holistic way. If you only look at the face, you can maybe understand most of it, but not everything. If we listen also to their voice, their tone of voice, then we have a richer understanding. So we decided to create two independent AI models, one computer vision, facial expressions, body language, and the other tone analysis. They work totally in parallel, separately, and every few seconds they combined. Oh, I think he's sincere. I also think he's sincere. Bing, we combined. So we have two independent points of views on a person which are more likely to be giving a highly accurate result. It's still a work in progress. We'll continue, of course, like everything. We continue to improve the data set. We continue to optimize the model for precision, but also for performance. People want results instantly or near instantly. So it continues to be a challenge, but the hard part of building the foundation, the data set, and the first model was done. And now we are with a product that is pre-commercial, that is already live and that we are piloting with different type of enterprise customers. We have 800 users more or less now as of this week. Adil Saleh 10:02 Oh, very interesting. And like you mentioned, you've done the hard yards. It's just about building the specialized layer on top as you get more use cases, as you get more customer feedback experiences and all that. I would love for you to explain some of the cases you mentioned, some of the coaches. I know that a lot of these enterprise customers, they have coaches coming from outside or they have vendors that actually help them get training and everything be done for their sales, marketing, product, technology, and everything. There are lots of consultants, I think 10 to 15 more, more than 10 consultants. OpenAI has, even from the beginning, they always had consultancy firms. So if you just tell us a little bit about this use case, some of top three use cases of these consulting or coaching customers for our deeper understanding for people listening as well. Chandra DeKeyser 11:01 Yeah, sure. Look, I would say the first use case. In companies of the knowledge economy, so white-collar kind of companies, it's talent war. The companies succeed if they are able to attract, nurture, retain top talents. And if you look at the poaching between Oracle, Apple, Meta, Microsoft, Google, they all try to steal the best talents, OpenAI, Tesla, they all try to steal the talent. So people. Adil Saleh 11:34 They stole some researchers recently. Chandra DeKeyser 11:37 Sorry? Adil Saleh 11:38 They stole some researchers recently, Facebook. Chandra DeKeyser 11:42 Yeah, I haven't followed that, but when people leave, they usually take their brain with them. So it's full of things that they learn in the competitor. So it's very delicate, this thing of stealing trade secrets, but you cannot reboot a brain and say, forget what you learned at Google and now welcome to Microsoft. But why do people leave? They are all well-paid everywhere, but why do they leave? Usually the main reason why people leave is that they leave because they have, just on top of them, a manager that is not inspiring. Whether they are 6.0 level or two, three, four layers down, you need an inspiring leader. And to be inspiring, that person needs to be authentic, sincere. There is a lot of politics and lie. I've lived in the corporate world. Lots of nasty tricks there. So this thing of authenticity, we are focused on that because this is key for companies to retain their talents. I like a saying from a coaching company in Silicon Valley, they say people leave managers, they don't leave companies. They may love the company as a whole, but right on top of them, there is this, I wouldn't call asshole, but someone that is totally uninspiring, taking credit for whatever they do and so on. So people leave. And so the challenge for companies is to coach their managers to be better managers and to be able to motivate, inspire so that people, when they wake up, they are really motivated to make an impact and deliver something valuable for the company. So this thing of making people better managers, more empathic. So this is a lot of soft skill. It's not a mathematical black or white. There's a lot of subtlety there. And it will depend on personality. So some people are tougher, some people are softer. You need to find the right balance. It may be a different balance for different people. So how do you help people become better managers? So that's a lot of coaching is involved. And whether the coaching is done internally or through video platform over Zoom or Teams or contractors come in and do workshops, there are all these kinds of methods, but in the end, they are all very human centric, but there is not a lot of AI intelligence in that. So how do you analyze how authentic someone is when they talk to their group of tenses, leaders? How do you do that in an objective way, unbiased way? So that's where we come in. We bring, our tool is able to analyze facial expressions, tone of voice in an objective way, unbiased. Why unbiased? Because we have trained our AI model on black, white, Latino, brown, yellow, all cultures, all ethnicities, all languages, ages, and genders. So we have an AI tool that is unbiased. It could always be perfected, but the foundation is really unbiased. I'm not saying that people are biased, but we have bias anyway, and we have good days, we have bad days. So this thing of judging someone, to help them identifying their weak moments in sincerity, authenticity, nothing better than objective tool to support the coaches. Coaches are excellent usually, and they have soft skills, empathy, but it's a tool that is gonna objectivize everything and do it at scale for them. So that's where we come in. And I was in sales for a German company way back. I was part of a group called the Top 60 of a company of a few dozen people that was quite something. And they brought this American sales coach. And he was talking about the salespeople, simplifying A, B, and C. They are always overachieving. They don't need a lot of support. They will go and get the information of the tech department marketing, and they will close the deals. They do the PowerPoint, they overachieve with the A, whether they overachieve by 10, 20, or 100%. And then you have the B in the middle that are on target, but they need to follow up on them, track them and motivate them and help them in the pitch and all that. But they are on track. And then you have the C that are, and so the common wisdom is you fire the C. But what you really should do is fire also the B because it's very demotivating for the A to see the B getting their paycheck and all that stuff. Now, rather than firing, which is expensive, often you have invested a lot in training, sales enablement and that. How about coaching the Cs to become Bs and the Bs to become A? So that's the approach. It's a constructive approach rather than kick out of the door approach. And it makes sense economically, obviously. When you have people that after six months, they don't deliver their quarterly target. They don't deliver. You could fire them, of course, but much smarter would be to give them a chance to elevate themselves. And that's where we come in. And it's huge need. I mean, all companies are stressed to reach their sales targets. So this is where the money is for us. And nobody's doing that at scale today with this kind of neuroscience AI. Adil Saleh 17:16 Love it. I mean, I think neuroscience AI is also going to expand as a category. I know it is not a category as yet, but of course, some of the companies moving in the direction of people management, people teams, they're playing a big part in feeding the addressable market. And a lot of these SMBs and mid-market businesses, they don't think of hiring outside kind of consultant or coach. They are trying to be coached. They're asking their managers or maybe founders or leadership team. They're sitting with the managers every month, every quarter, and they're giving them, hey, these are the mistakes. This is how we can mitigate these. This is the best action plan. This is how you can get best work out of people. So do you think there is a problem of adoption into these platforms? How do you see this as a leader? Chandra DeKeyser 18:20 Adoption of what? Adil Saleh 18:20 Adoption of tools like these, investing into tools like these. Chandra DeKeyser 18:26 Look, it's not a question of showing the ROI. And everybody has a sales target that is hard to reach. And when you reach the investors, the board will ask for more. You know it when a company raises series A or B. Someone will think, oh, game over, series B. No, series B, they breathe in your neck and they fire your ass, including CEO or the founders, if the sales don't grow exponentially. So the sales, the need for companies to grow is always there. And it's very stressful. Of course, the top executives make a lot of money, but their head can roll very quickly. You miss a quarter, you're still alive. But you miss two, you're dead. And it was your company, you built it from the ground up for 10 years. It's not personal, it's a business, they fire you. So the ROI is compelling because we are able to spot. If a sales pitch takes 10 minutes, of course, a good salesman will also be a good listener. But let's say there are moments when he or she's talking and you want to see automatically. We analyze the sincerity every second. So every second we have a sincerity score. And then when it's falling off the cliff, then you know there is a problem. And you can pinpoint the salesperson, hey, that's where you're failing. That's where you're losing your customer, your audience. Having that automatically for salespeople to rehearse from home, or their home office or whatever. And then having the sales leader being able to, show me the moments of weakness of my sales team so that I can help them, or show me the ones who are high performance because I can maybe use their pitches as an example for the others. Doing that automatically at scale without adding one person in headcount. This is huge. And so initially, we were not too good at showing the ROI. Now we have brought an amazing person. I'm very proud of myself because I'm a sales guy and I'm selling the company to other people. So now we have a new board member and he was the Chief Revenue Officer of Amazon Web Services. So billion dollar deals. He was chief revenue officer also of Oracle way back, still knows personally Larry Ellison. So huge track record, also other companies, but these are the two main companies where he was CRO, chief revenue. Now he was playing golf and investing a little bit. He joined our board and he's helping us also formulate our value proposition. Because one thing is the product, what the product does. We need to express it in a compelling way. And that's the art of sales. For each market segment, we need to be able to formulate this in a very concrete way. What am I going to do? Close rate, you're going to increase your close rate by 40%. You're going to increase your revenue by 25%. And it's going to cost you 10% of what you're going to earn. So the ROI is compelling. It's not like a year ROI. It's like three month ROI. Adil Saleh 21:42 Yes, so it has to be tangible, absolutely tangible and something you can measure across the journey. So now getting back to a little bit on the product and application. I know that you already mentioned the sweetest segment is more like consultant coaches or leaders within the organization. So now thinking about these segments, how they're able to get into the platform. Could you walk us through a little bit about their experience, how you're measuring success for your platform? I know you mentioned that you have around 500 plus customers already sitting, consuming the platform. And whatever they need to help, what kind of customer education initiatives have you guys taken internally? I know that some years is quite a long time in this AI evolution, all of this. You can do a whole bunch of education guide, knowledge base, sort of success metrics around triggers, how you can measure success, how you can help them expand, a number of people, what kind of expansion plans you have. Chandra DeKeyser 22:52 Firstly, let's talk about the number one segment on which we are. So it's the sales leaders. So there the ROI is compelling. It's really the KPI for success, how we measure success, by looking at the closing rate of the salespeople. So it's really impacting the top line of the company. They start selling more when they use AI to coach their sales force. We have other customers in the field of HR. They are compelling, but maybe let's talk a bit more about the sales leadership. So we've just finalized a pilot with a customer. So they were coaches helping sales reps or sales executives deliver more compelling presentations to customers. So what they were doing, they were filming the salesperson who was on stage in front of a group. So it was an in-person workshop of two days of sales leadership. They were filming them at the beginning, in the middle, and after two days at the end. And during the two days, they were being coached on different aspects of how to deliver a message, the body language, online. So they were filmed three times. And then the videos were analyzed by our tool and you see a sincerity score aggregate, and you see the sincerity score growing one, two, three. Out of 20 people trained, I think 17 showed an improvement. So it's quite amazing that you can measure it with numbers. You can put numbers because usually you coach people, they go on stage, they pitch. You have no objective data coming out of that. What data? Nobody's giving data. You can write a story. That's a four, that's a five. Totally subjective, wet finger. No, this is AI with algorithms that are trained on analyzing facial expressions, tone of voice, that give numerical values every second for every of these pitches. Then you aggregate these numbers and you see sincerity, sincerity. And the third one, sincerity is there. So it's growing, it's improved. So you have, and for the coach it's very interesting because they said, because they have to send their services to the company, they say, look, with my methodology, with this tool, I can demonstrate that the sincerity of your sales executive is growing. In two days! Now they'll go back to the front line selling to customers. We'll see in two months, three months, how many deals they close, how does it compare? And then with a bit of statistical analysis, we can show the ROI and they will see how many more, the cashflow changing. So that's the most exciting for us because companies are self-driven. Now, the other one is the talents. The talent war is very important. A lot of people are applying for a lot of jobs. We have a new project, a product in the R&D stage, almost finished. It's a fake, a deepfake detector. It's now interview, job interviews. You and I, okay, I know you're a human. You probably know that I'm a human, but in two, three, four years, are you going to be a human or an avatar that is lifelike? Deepfakes are coming everywhere. Now it's a chatbot. When you're selling on LinkedIn or email, you get all these messages that are AI generated. It's an LLM that is creating all these messages, right? And the message is personalized by looking at my LinkedIn profile. They will say something to make it feel that it's unique to me. So that's all LLM text messages, chatbots, email, LinkedIn messages. Now, it's going to go video at some point because it creates a more personal connection. So anyway, for the job market, more and more people are applying and they are fake. They are applying maybe during the video call, someone appears, you don't even know if that's the real person you're hiring. So the deepfake is becoming a problem. Gartner made a study that in 2030, which is around the corner, time flies, 25% of the job interviews will be with a deepfake. In five years, 25%, one out of four people that you interview for a job will be a deepfake. And how will you be sure it's a human that you're interviewing? So we're working on an AI model to detect that. And we already have this AI model for sincerity. When you are asking people about, one thing is the technical skill. You don't need AI to analyze the facial expression. You want to see if the guy or girl is able to answer a few technical questions and maybe they have chat GPT on their other device to give the answer. But let's say you do a technical, but there are soft skills. The motivation of a candidate to join my company, for me, it's very important. I want to see if they are truthful about their motivation. So now we eat our own dog food. We ask them to open an account on sincerity, to create a video selfie, five minutes maximum, where they talk about why they want to join, why us, why me, why now? And they send it to us. And we see the sincerity scores. And so we have a big bank in Colombia, biggest bank in Colombia. They are now using the tool to screen candidates. And so big banks, they hire thousands of people every year. And for them, the cost of hiring is expensive. They go to three, four, five interviews of one hour each. It's a very costly, it's a big department for the bank, HR. It's going to be probably reduced with AI, of course. And here is the first step. It's using AI to see if the candidates are truthful about their motivation to join the company. Then you can have 10 people going through an interview. Then you push a button, you get the report. So the graph of each of the 10, then you can see the top three, the bottom three, and the bottom three, unless you really like them for a specific reason. Adil Saleh 29:04 Are you mentioning like they're doing 10 interviews at once? Chandra DeKeyser 29:07 No. They are interviewing one after the other, but then all these videos of interviews are being uploaded to our platform. Adil Saleh 29:14 Okay. Chandra DeKeyser 29:14 And automatically we analyze, we compare the candidates for that. Adil Saleh 29:21 Oh, do the comparison qualification-wise and motivation-wise, all of that? Chandra DeKeyser 29:25 Yeah, motivation, not qualification. So that's, we don't add value for it. Adil Saleh 29:29 Okay. Because a lot of other factors also come into play, like apart from sincerity and motivation, all of that. So do you also compensate all of those data points and metrics? Chandra DeKeyser 29:39 Oh, look, I mean, we are only bringing data on the sincerity. There are many other criteria for a bank to hire people, and we are not intervening there. But this step was not present unless you had some psychological evaluation. But again, subjective, and a psychologist, even a well-trained psychologist may have a good day or a bad day. So this is objectivizing this step and automating it. So you can also reduce the headcount for the interview process, which is expensive, to have objective data at lower cost. Adil Saleh 30:12 Very interesting. And now, thinking about how you are measuring as a marketing team, how big is the marketing team that you have? Marketing or post-sales? GTM post-sales. Chandra DeKeyser 30:25 Right now, we're still small. We're less than 10 people in the company. We are hiring a new COO this week. So we'll be nine people at the end of the week. We have one marketing director, and one junior, deputy. And we have a Chief Growth O fficer who is based in Austin, who is a seasoned marketer working on driving that. So it's three people, one of which is working half-time. Adil Saleh 30:53 Okay, interesting. And now thinking about how you are, of course, do you think that it's kind of a white-glove service, that it's fine for the enterprise because the contract value, lifetime value of the customer is big enough? If you are to enable SMBs and mid-market customers, you're in the value, right? You know that startup ecosystem has this problem. A lot of people during the hiring lied to them. They figured out three months later that they made the wrong decision. That cost them a lot. Same as Austin, Texas, and San Antonio, they're going to be big in tech as well, New York City. So what kind of measures and initiatives are you guys taking towards making it at least 60% to 70% self-serve so you can deliver value on a self-serve basis, and then you can go upmarket with the SMB and mid-market customers at scale? Chandra DeKeyser 31:46 No, the tool is already self-service. So it's a web app to upload videos. You can do it manually. You can do it automatically with a kind of scripting tool that automates the upload of hundreds of files. And we have an API that is going to go live in September, end of September. So with the API, you can have basically any system communicating with our platform. And we also have an on-premises version that's a bit more sophisticated. We have some onboarding, installation, configuration, but we can install our platform to the cloud platform of the customer so that their videos, which are often sensitive. Adil Saleh 32:23 You're actually living in their environments. Yeah. Deploying in their environments. Okay. Perfect. And that is probably more for these bigger logos companies. Chandra DeKeyser 32:33 Yeah, bigger logo, but most even mid-market. Our sweet spot right now is companies with a few hundred employees to a few thousand. We're not going yet to the top. I mean, this bank in Colombia is a bit of an exception. We will, of course, welcome any customer to our door, but when we go outreach, we try to go to companies which have a few hundred employees. Why? Because then they have probably 20% to 30% in sales, 20, 30 salespeople as a minimum, and then they can see really tangible value. Sales, B2B companies, obviously. So salespeople who have a target of maybe a $1 million ballpark figure, sometimes a bit more. Let's say 20 salespeople with a 1 million. So it's a 20 million sales target for that company. That's a good spot for us because it's easy to talk to the CEO, to the CFO, to the CRO. These are approachable people, not like in a Fortune 500 where you have to go through kind of a. So it's easier to talk to the key decision makers. And we've started doing that over the last few weeks. Quite interesting feedback. Adil Saleh 33:47 Interesting feedback. Perfect. Love that. Chandra, it was really nice getting to know Voicera and getting to know you as a person and the kind of vision that you have. I mean, we keep on following your journey and definitely wish you best of luck for all that you guys are doing because this problem has a critical nature to it. It's about people. It's about the sincerity of people, and it's about living a business impact that is, again, led by people. So I really love the way that you're approaching this application, this solution. And I wish you really good luck towards doing it at scale in the mid-market and achieving big success. Chandra DeKeyser 34:31 Fantastic. Adil, thank you so much. Trust is essential, and we have been measuring it with the gut feeling that it's sometimes good, sometimes not so good. We've been lied to a few times in our lives. Even if you're a millionaire, I'm sure some people have lied to you and it affected your decisions. It happens all the time. So having a tool to objectify that, as a tool to help you take the better decision. We're not taking the decision. We're giving data to help people take decisions, using them or not, or the way they want. Adil Saleh 35:03 Okay, perfect. Yeah, it was a great conversation. Chandra, take good care of yourself. Have a good rest of your day. Chandra DeKeyser 35:10 Thanks for having me, Adil. Take care. Outro 35:12 Thank you very much for listening. If you got one useful GTM idea out of this show today, please share this with a teammate and hit follow. Explore Hyperengage at hyperengage.io and Dextego at Dextego.com