Brainfood Live On Air - Ep398 - Meet Ema - Your AI Employee for HR, IT and Finance (Live Demo)

0:06Yeah, definitely. I mean, don't do anything that causes us, you know, to expose data we don't want, etc. All of that is no good. But whatever we can see that you're prepared to share would be fantastic. So good stuff. Okay, we're live, everybody. Welcome here to Brain Food Live on Air, bringing it to you, I would say every Friday, but we have a midweek special today. And the midweek special is always the same. It is kind of our challenge to all of these wonderful new products that are emerging in the market today. My challenge to them is to say, hey, listen, you're telling us you've got this amazing product.

0:42We would love to see it live. And I'm delighted to say a lot of founders are very confident and they're very happy to come and share what they have. And today we have one of those founders. It is Surajit Chatterjee of EMMA. Surajit, wonderful to have you on the show. Welcome to Brave Food Life. Very excited to be here. Thank you, Han. Great stuff. Surajit, by the way, I'm going to share Surajit's LinkedIn on the chat stream here. And if you don't know Surajit, you certainly should. It's got to be one of the most impressive CVs out there on LinkedIn as we see it. You work for Oracle, you work for Google, you're a CPU or Coinbase.

1:23I mean, this is a rarefied atmosphere when it comes down to working for these big tech organizations. I wonder whether you could tell us a little bit about that trajectory and what it's taught you when it comes down to launching a product today. What is the sort of being around some of these really profound technology businesses? What has it taught you about how to build a technology business in 2026? Absolutely. And thank you for that generous welcome and kind words. Look, I spent last over more than two decades in the Valley, as you said, working for Oracle and then working for Google, where I had this opportunity to actually see a company, and it was still early days, 2007.

2:13I started working on mobile advertising at Google, and nobody was working on mobile ads. So it was like me and a couple of engineers, a really small team, starting it out even before smartphones kind of came in so um and and it was great to see a company taking like giving that leeway to someone who was at the really at the at the base like the you know very entry level i joined google as an entry-level product manager so uh but mobile ads turned out to be the right bet for me and for google um and it when i left google it was a hundred billion dollar plus revenue per year so i had got promoted i don't know 12 or 13 times i was running this massive team uh global team uh running mobile ads and then google shopping google adsense for search more than half of google's revenue um at us at some point so that was an incredible experience to see that kind of zero to one to kind of literally 100 growth.

3:24And what innovation can do, what risk-taking can do. Because when I started working on mobile, nobody thought actually mobile will be that big. In fact, nobody thought mobile ads or mobile commerce will be that big. Now it seems like very obvious, but it was not obvious at that point. In fact, this funny anecdote, I remember my sales team, Google sales team at the time would give me like five minutes at the end of like a sales meeting because I'm this little mobile person on the side and they have to sell everything else. And I was pitching to the CMO of Macy's and he was like, this is great. I like your enthusiasm, but nobody will buy a dress on a mobile phone.

4:10And that's not going to happen. People want to come and touch and heal and buy the dress. And we all know where it turned out. We buy everything on mobile or online. My next phase of my journey was at Coinbase. I got a call from Coinbase's CEO, Brian Armstrong. One day, it was kind of serendipitous and just out of the ordinary. and he wanted someone to come in and help scale coinbase uh it was doing hundreds of millions of dollars of revenue i think 150 or so at that time and my job was to really build out the product and engineering team diversify the product portfolio and grow canvas into a multi-billion dollar company and that's that's what happened next few years we went public um in 2021 and then 23, I decided to step out and start this company.

5:09I felt I wanted to build more and have built for larger companies or mid-sized companies and wanted to just go at the basics again and start from scratch. Yeah, fantastic. I mean, it's crazy that you've seen these huge technology transformational moments um i mean working on mobile ads pre-smartphone um i think there's probably a generation of people that don't even understand that that little period where there was a smartphone there was a decade where there were mobile devices that they weren't smart um and it was just text based we were filling in with hard keys um and it was inconceivable at that point given the size of screen that you would actually ever do any transactions on it but of course big technology revolution occurred smartphones were ubiquitous and i do believe now i mean i think there's something like two billion about how many billions of devices are there it's probably more isn't it it's probably like 10 billion mobile devices around today isn't it i i don't have the exact number but yeah it is probably more than the number of people because a lot of people have multiple devices so yeah absolutely incredible game-changing and to be able to see and drive to a certain degree sort of develop that process is fascinating.

6:31With regards to the business you're running now, Surajit, and we're going to talk about this in focus, was there a particular moment that you realized, okay, it's me that has to do this rather than, hey, there's a product that's coming that will solve the problem I'm encountering? Or did you see kind of a little bit down the road, around the corner and think, you know what, there's an opportunity. should really grasp it yeah look i always try to kind of do things a little bit early in the cycle so you know when i worked on mobile it was very early in the cycle crypto not too early but still very early crypto was not mainstream at all when i joined coinbase uh with ai you know 2020 late 22 when chat gpt and everything else was coming out i got very excited thinking you know it's not just about generating words that's what ai was doing generating sentences phrases whatever predicting the next token can you actually teach ai to plan and act see if you can do that and then learn from that action right then it becomes more human-like and it can actually you know substitute human employees for a lot of the mundane and routine tasks that people have to do.

7:54And that's kind of my experience. I have been in these great companies, Google, Oracle, or Coinbase. I had great talent, but 80%, 90 % of the time is spent in just keeping the system on, right? Keeping the lights on. We used to call it KTLO at Google. Even at Google, sometimes 50 % of engineers' time will be just to keep the lights on. Right? Yeah. Another word to describe, another phrase is minding the machine. Like there is a system and you just got to somehow keep the system going, whether it's sending a bunch of emails, moving some documents here or there or not value add. It's just moving the thing through, but someone needs to do it without sort of, without this substitute that we're discussing now.

8:41So absolutely. Yeah, so that was the inspiration. Look, I believe in kind of the spirit of us humans to build new things and create new things. I think everybody wants to feel great in the jobs they are in and add real value rather than copy paste things on a spreadsheet and so on, like do the same routine thing over and over again. So how do we free up time? How do we actually create AI employees that can take a lot of the routine tasks? So that was the inspiration. In fact, my first pitch deck to VCs had the award AI employees. In fact, nobody understood, to be honest, nobody understood what I was talking about.

9:29I was I started tinkering on my own reading a lot of the research papers in fact connecting with many of the PhD students they were all excited by the way some of them started working with me and like pro bono because we didn't even have a company but they were just excited to to see that someone wants to build a real product on these concepts and that was the genesis of the company. We built the platform to build AI employees that can actually reason, act, have memory, can respond to human feedback, can learn from human feedback. And of course, the whole industry has now embraced this verbiage.

10:13You see the AI employee verbiage everywhere now. Funny story. When we were coming out of stealth, we built the product almost for 14 months or so. We didn't even have a website. We didn't build one. We wanted to build a good product first. And when we are coming out, we had a PR agency for advice or whatever. And they said, no, no, you should remove the word AI employee. That's very provocative. And so they scrubbed, forced us to scrub everything out. And we were two days before the launch. Like I just didn't feel good. So I called them and said, hey, this is what I want to be true to what I am building, true to myself.

11:01So I went and changed everything back to AI employee. We called it an AI employee, right? And I didn't feel that it is actually that provocative. Maybe it may seem provocative, but I think the future is human and AI working together. Right? That's the future. Yeah, that's it. It's the collaboration. I think if we kind of reframe the conversation from AI versus human and think of it as human plus AI is the optimal solution, if we all kind of look at that as the North Star and work towards that in good faith, then I think you'll get a lot more support. And I think, you know, it's early days, Surajit, but I think your intuition to go back to AI employee is getting validated.

11:56I think we're getting more comfortable with the concept of these AI agents operating much more like colleagues than they are tools. You know, they're not stuff that we pick up and just put down. They can autonomously operate. they can remember, they can learn and we hope they can execute that they can take on the work and that's one of, go ahead, Sujit Sorry to interrupt and also very interesting, you might have seen the news they respond to encouragement like humans AI just called math problems So the person just encourages AI that, oh, you can do it, which is fascinating to me, right? That's a very human quality.

12:47And, you know, you just opened up a door there, Surajit, and we have to just step through it, because I think it's too exciting not to explore just for a moment, guys. Perhaps the skill of the manager then will also evolve, because it's no longer just about like cultivating, developing human employees, but presumably it's also about cultivating and developing AI employees because again, if they are genuinely agentic in the sense they can learn, their performance can improve, presumably their performance can degrade also. Presumably their output will go up or down in a variable manner based on their experience of how it is.

13:26So it'd be very interesting to see how this creates a demand for managers to develop this skill set of being able to work patiently and inspirationally perhaps with AI employees? Yeah, absolutely. I think just to give you one data point here, one of our customers, he is a chief experience officer at Wipro Technologies, the largest professional services firm in the world. They have 250 ,000 employees and they use MR for entire hire to retire cycle. Right. So he was telling me I was asking, so, you know, how is your organization evolving? And his comment was, look, I think my entire team has to retrain because everyone is now a manager.

14:18They are managing AI employees. It is a skill set. So even people who are earlier in their career now have to learn how to manage and how to manage AI. So to your point, this is already happening and true. And it's going to be a very interesting world going forward. Interesting indeed, Surajit. And before we get into the presentation, you mentioned that Emma already in with Vipro. everyone knows Vipro, a huge organization. Can you give us a quick status update as to where Emma is right now as a business and also as a, you know, who's using it and where is it deployed? Yeah, so we work with Fortune 2000, largest enterprises in the world.

15:10NTT Data is a client of ours. Many other Fortune, like top five insurance companies, top five automotive company, and the largest BBM in the country and so on, pharmacy benefit management, healthcare providers. We built a framework that can automate business processes, any type of business process, but we have seen most success in HR IT finance, like employee experience end to end. So today I'll talk about recruiter, which is the beginning of the employee experience. You can think of sourcing, recruiting, onboarding, But we work with companies like Wipro and many others deploying this end-to-end platform where there's one-stop shop, basically one single pane of glass.

15:59It could be your Microsoft Teams or Slack or our own UI where you can get anything done, everything and anything done, whether it's onboarding a new employee, recruiting employees, performance management, learning and development. any action, you want to change something in payroll. The biggest thing we are solving for is today every enterprise is a jungle. It's with hundreds of thousands, sometimes 2 ,000 plus applications in large enterprises like SaaS apps. Nobody knows where to go and do what. And not only that, every application is a silo. It's an island of its own. It has its own data, its own little UI.

16:44It works great, but most work gets done in between those applications, not just inside those applications. Today, humans are the glue between these apps and the data sets you have. We are removing the human glue so humans can do more value-added work, more strategic work, rather than going copy-pasting data from one place to another or doing manual workflows and so on. So that's the big kind of ROI. We had deployed at massive scale, like in Wipro scale. We are doing like 50 to 100K requests every week, resolving them. They report almost 50 % reduction of their HR ops team that they have moved to other areas and so on.

17:33And their employee satisfaction has grown over 20 % point during that time. They are doing over 100 use cases already and we are doing a lot more. Similar story in Hitachi, one of our largest customers, same thing, deployed at massive scale with multiple Hitachi companies and using us for tens of different use cases. So today we are going to deep dive into one use case that is very dear to me, which is sourcing and recruiting as a founder. that's a very important task I have to do and frankly I couldn't find anything out there that's actually reduces my time like you have to hop between tens of different applications it's like what you call the tool tax you have all the tools but you have too many of them yeah how do you streamline the process and that's what I'll talk about a little bit Yeah, fantastic stuff.

18:29Good to know where Emma is in terms of the distribution and the significance that it's already live in production with these huge organizations is great. Also a bit of an indication as to how, shall we say, deployable it is. You know, we're not talking about, you know, a three year sort of implementation plan. It looks like you can just get this into the system very, very quickly. That's the biggest thing. Look, traditional SaaS sometimes takes like nine months, a year, two years to deploy at a large scale for us, like the entire Hitachi deployment, which is across like 30 ,000 employees or 40 ,000 employees, across multiple companies, because they were trying to build a single back office for HR across multiple holding companies.

19:25It was just eight weeks because our AI employees, we didn't need to move, they didn't need to move the data or consolidate any system of records. Our AI employees just like acting like humans, reached out to the right data sets in real time and get the job done on behalf of humans. For some of the products, the onboarding is instantaneous. I'll show you Recruiter today. There is really a couple of things. No onboarding. You can just come and use it right away. The vision I'm getting here, Surajit, is that basically these big corporates have just acquired an infrastructure, an architecture, if you like, that's cobbled together over time.

20:11That's just very normal, but it's now so complicated that you don't want to break it apart. You don't want to need to necessarily consolidate it. You just put agents into the system and the agents just navigate through this sort of convoluted structure and then transport the data where you need it to go. And I think everyone can kind of visualize how that makes perfect sense because you retain the existing system such that you don't have to do a demolition job before you do a rebuild. Absolutely. And that's the big change in this transformation this time. There's an earlier time you have to kind of get rid of some application, move everything to a new application.

20:55Those transitions are very, very painful for any enterprise. Now you have the time. You will probably need to eventually realize that some of the applications are not needed and you can slowly deprecate them. or remove some of the licenses and so on. That is also happening. We do think over time, some of the SaaS will die, their natural death. But this is it. It's a natural death. It'll atrophy away through lack of use rather than surgical intervention, which causes huge amount of damage. And then there's a recovery process with the organization that you've just ripped this thing out. And sometimes companies never recover in the same way.

21:37So I love this concept. Okay, Surajit, let's see what we've got. I mean, I think you're going to show us a few high-level things to begin with. So I think people would appreciate that. That's fantastic. Yeah. Let me pull up a couple of slides. I promise you we'll get to the product real soon. If you can see my slide deck, hopefully. Yeah. Yep, that's perfect. Okay, cool. Awesome. Awesome. I think you already introduced me, so I'm not going to go through my introduction again. And as I talked about, we work with some of the largest enterprises in the world, and this is just a sample set of some of the companies we work with.

22:19And you can see the logos and see how large and scaled they are. And it's a different skill set to work with large enterprises and being able to demonstrate value. And in every one of them, we're demonstrating value. This is a quick case study with actually coming to recruiter. And specifically, Artico is actually an executive recruiting firm. And Mercedes, she is the CEO of Artico. They started using our product from the very beginning. Actually, they were the design partners almost working with us hand in hand as we built the product. And they are seeing a tremendous impact, like 70 % saving in time to hire, 60 % of their volume completely handled today, 30 % less cost for hiring.

23:08And you will see in the product why that is the case. And this is what recruiters or HR teams spend their time today. They spend their time on tools, not talent. And this is not only for hiring, by the way, I'll say even internally, once you hire the talent, the HR team is spending all the time on tools rather than actually thinking about how to grow that talent internally. And we have employees for all of those use cases as well. We go very deep into HR and IT use cases. So, you know, they are juggling with tools, as I said. And the other thing is sometimes some of the best candidates may slip through because recruiters may miss real signals because there is so much data that is available for you to kind of shift through.

23:57It's humanly impossible, right? So hiring costs kind of keep rising because of the tool mess and the data, the challenge in accessing data, consolidating that data and bringing it all together. And that's what we have solved. So you will see in the product, you can focus on talent, focus rather than on the tool. All of your data can come together very seamlessly, simple conversational interface. and we have our own data set. So we work with 30 different data providers. We bring in like 1 billion plus candidates and we have very deep intelligence on those candidates. But it's not only just selecting the candidate, it's the entire process.

24:42Today, if you look at it, the steps in recruiting is like position management. You have to create a job description and so on, upload it somewhere, publish it, sourcing, ranking, outreach to the candidates, because you may like some candidates, they may not respond to you. Sometimes the best candidates are not applying. They are not really actually looking for jobs, but maybe to reach out, they may be excited. Then you have to do some pre-screening, scoring, onboarding. But if you look at them, this is an entire workflow. Thinking of this as a single tool is is probably the challenge that has existed in the past so the goal of our ai recruiter is automating every steps in hiring all channels seamlessly into one platform you'll see that email linkedin uh everything else all in one platform uh into massive intelligence at scale right, billions plus profiles, multi-channel outreach, role-specific training, and the last but not the least, governance by design.

25:55This is why largest enterprises in the world actually choose EMR because we are most compliant startup. I can confidently say that if you go to trust.emr.ai, you can see our breadth and depth of compliance and security. No PII ever goes to a public model, will automatically redact them human in the loop when you need that because not everything should be fully automated we completely believe that and human judgment is needed i'm going to skip through this very quickly this is just showing all the different types of use cases that's embedded into the journey of a of a recruiter and i'm going to quickly flip over to the product itself um let me just try to do a quick uh quick search on myself and and demonstrate what you can do so i am one sec uh so just we're no longer looking at a slide deck um this this is at least a sandbox or maybe even a live site um but it is the application um and if i'm a customer of emma Let's say I have access to this.

27:06This will be sort of the website. And I will be able to look at this screen and do something similar. Yes? Yeah, absolutely. Absolutely. So if you just go to recruiter.emma.ai, you can log in on your own. And you can create an ID, et cetera. It's very self-serve. You can just go through it. Make sense? Perfect. Cool. So what I'll do, look, we are actually right now growing our sales team and we are trying to hire a lot of salespeople. So, you know, I am actively looking to hire sales. Right. So the first thing you can do, you can point Emma to a JD. So I can actually point to a JD. I can actually, if I flip over here, I can give it a JD and it will take the JD and start working on it.

28:07For the sake of time, I'll not upload a JD. I'll just do this search kind of in real time, right? So I'll just type in, you know, Anu Khaer and AE with experience selling in healthcare industry. I am cheating a little bit. I wrote down some of the prompts so I don't have to type in and spend the time, but I can also type in anything. And I'll just ask it to do that. Now, here, it's really going through this 1 billion profiles, 30-plus data sources, and you can see how quickly it comes up. And it goes through, and it can also pull in data from your internal ATS. So most companies actually that we work with, large enterprises or medium to large size enterprises, already have hundreds of candidates in there or thousands of candidates in their own rating.

29:00Yeah. Like people who already applied before, right? And many of them may be good candidates. So we are not only bringing in information from outside. We are also looking at internal talent. It could also be some current employees who could potentially apply for the job and so on. So in both internal and external talent, and it basically looked at, it ranked everything. You can see that it did, you know, 776 ,000. So it profiles it within seconds, as I was speaking, this is all live. It actually ranked and created scorecards and so on, right? that's that's the speed at which you need to work now i want to actually refine my search a little bit right so you know i'm like okay this a should sell our software um for example it's saying you know about seven seven hundred and seventy six candidates match this search the search itself is complete but do you want me to tighten this search if you see my screen here at the bottom This is our autopilot.

30:07You are basically conversing with autopilot throughout the entire process. So I'm going to tell it, yeah, I want to tighten this search. I want to further refine it. So say they will sell our software products through this industry. Let's discuss the companies we should look to target here. So I'm not actually put anything in here. I'm saying I want to discuss some companies or some verticals that I should target. And it's to think through, okay, what vertical choices should it give? So it's giving me a bunch of choices. Okay, maybe EHR, healthcare, pharma, because within healthcare, there are many sub-verticals that I can go for, right?

30:52And it's asking, which one do you want? So I'm saying, let's focus on vertical-focused company. So I'll say it's saying, should I focus on a vertical focus company? I said, sure. Let's focus on vertical focus companies. I can also say focus on like specific healthcare data or specific vertical as well. It's going to update, right? It has added 22 healthcare vertical SaaS companies as a nice to have boost. and it further refined the search, right? I'm going to ask it, okay, I need this AE to sell on the West Coast. And as you can see, each one of these would have taken me a long time in previously going through filters, doing all kinds of stuff and so on.

31:50I'm just here just chatting, conversing with it, saying, okay, I need this AE to be able to sell to the West Coast. So it's going to further narrow based on location and based on where the resume is, where it says they have sold to the West Coast and so on. So it is further narrowing. Let me just now say, okay, maybe I forgot to mention something. I forgot to mention that I'm looking for an experience to eat, maybe five plus a year is probably. So it's going to further go ahead and refine and say, okay, I'm going to get further refine and get an experienced AE, which is who has five plus years of experience.

32:45Now, all that is external search. I can now go in and say, okay, can you look at internal search, for example? Can you look at which internal talent is working most closely with healthcare customers right now? So it's going through my internal talent base and saying, okay, I found someone, Jay Rockett. He is a large enterprise account executive. He has worked at healthcare at workday before. So if I go and find here is Jay, I can search him up. And I can go deeper into the candidate. And you can see this. It has pulled in all kinds of data it can find. Public data, data from data sources we have, its entire profile, the scorecard.

33:42Again, based on all the things I have asked it so far, it has actually rated Jay and put in a little bit of commentary like why it thinks like it's jay is 4.7 overall uh based on the profile and and data i'll not click on contact info for privacy reasons it also has ats notes so this is like past interviews we did with jay for example it has pulled in from our own ats It can also go in and look at a lot more detailed conversations that we have had. And this is from our own, say, Ashby, which is one of the ATSs we use, and pulled in information. So it's pretty detailed. Everything at your fingertips, all the information is here.

34:36and something that would have taken me probably hours just done very quickly. It forks off multiple agents in the background and does all this work kind of instantly shifting through literally millions of profiles. Yeah, and I think this is where the sort of the breakthrough moment really has been for this generation of artificial intelligence. They are large language models, so they are ideally optimized to pile through these huge amounts of data, which obviously a human being cannot practically do or any team of human beings cannot practically do. And certainly not as effective. A couple of questions coming in is like, OK, where does the data sourcing?

35:23You mentioned 30 plus public data sources. So we can kind of infer what they are. I guess it's LinkedIn, it's Reddit, it's anywhere where people are publishing. do you kind of do you hold a candidate record in a system or is it kind of generated real time is it a case where I've heard people say hey the database year is over literally we can just generate it right here and we're never going to hold the data we're just going to real time deliver it so look we actually license data from many proprietary data sources So not all of this is public data or publicly available data. It's publicly available data licensed through special arrangements, many data providers across the world.

36:18And we also enrich that data using additional publicly available information and so on. One thing I didn't show you, for example, if you go to a candidate, you can look at, say, news. It's the news about sometimes not all candidates will have kind of news about themselves, but it will actually go and find news about their company, which may be interesting or useful to you as you maybe talk to the candidate, right? So maybe the company is getting bought over or getting acquired by a private equity firm. So you are having a conversation or some event has happened in the company that may help you understand, or maybe this candidate is open to switching.

37:02Maybe there is a leadership change that has happened. But a quick question, Surajit, could that be part of the search? like you could say hey um i particularly want to identify candidates that have uh their employer has been prominently in the news or uh uh over the last three months or so something like this yeah you can type in anything you can type in anything in this uh conversation and it will go and do it for example here the next thing i i will do here is look i am looking for someone who can manage a team. So looking for someone who can manage a team. And it will actually go back and do that.

37:47So all the information is searchable and all that search, you can combine all that search together. And that's the power of it. We couldn't have done it before, frankly, two years back or three years back. You couldn't have done this at all. This is the power of the new technology. And the underlying, we have tens of different agents working together. We're spawning agents in real time, and there's a whole agency harness working together and doing this, using data that's proprietary and using data that's out in the public domain, combining all of that and getting it. So look at that. It has, and the other best thing is it's explaining everything it's doing.

38:30So it says, I have added team management as a primary criteria, right? And it also, what is interesting is I didn't talk about Jay Rakat anymore. Remember the internal candidate? It says, the internal candidate we just looked at has had several sales management roles at ADP. So we should score well on it. So it just on its own saying, oh, maybe you should look at the internal candidate also for this. even though I didn't explicitly ask. So I'll ask it. One interesting thing here is this explainability, right? So just for fun, I'll say, look, it seems like... So the top candidate here is Sabrina, who is from Carmel, California, or sorry, Carmel, India, Indiana.

39:24and the lowest rated here is Kevin for this search.

39:34I'm asking why is there a difference in your ranking? Again, look, AI is not making the final decision, so you can always decide to shortlist whoever you want to shortlist based on your judgment, additional information. You may have information about someone who worked with that person and so forth. So AI is just a tool. I just always want to make sure. But the interesting thing here is just explain, okay, for the criteria you have given me, this is why I have rated someone a little bit higher than someone else, right? While Sabrina is currently an A at Next Tech System. It's a healthcare, one of your target companies.

40:15And before that, she spent eight years at Compulink, which is another healthcare company, right? While the other person had less experience and then was in a security and home insurance kind of company. So that's why I rank differently. Seems reasonable explanation. But of course, again, as I said, the final judgment is with humans in any such situation, right? AI is just helping you do things faster. So what I can do now, I can shortlist a few candidates, right, because I can talk to everyone, and I can say, okay, let's shortlist your top five candidates. And I can go and click here and do that too.

41:11but i i just prefer to just tell it what to do

41:18yeah and we can see that's going to be the default way to interact with all software of this type it's going to be conversational um even the concept of you know drop down filter uh all the little sort of widgets that we used to have to learn over of a in a piece of software that will also atrophy um it will increasingly become just the chat interface absolutely so i can click so it created this short list list here so i can see okay it shortlisted the five candidates here right and i can go and now uh do a start an outreach to all of those candidates so i can say start an outreach so it will actually start building a sequence and and start outreaching so let me build a outreach sequence.

42:08So I'll continue the setup. So here, the big challenge in outreach today is, look, you have to go, you can do an email outreach, you can create an email sequence, or you can create a LinkedIn sequence by going into LinkedIn. But you cannot actually kind of combine the two and create a single sequence that does some LinkedIn, some email, and go back to LinkedIn and so forth. Right. With us, you can actually bring all of this together. So I can say, add an email. So it will create an intro email. Again, I have flexibility to change the email any way I want. I can type in and change. I can ask it to change and so on.

42:52I can say, okay, then add a LinkedIn connection. Right. I can ask it to say, add a LinkedIn email. And this is like what a multi-touch outreach will look like. right so i'll wait for and by default it's choosing two days to wait i'll wait for a couple of days uh maybe it's too small a font um and then i will send a linkedin connect request uh i'll wait for a few more days maybe send an email on linkedin i need to have the ability to send email and then i may send another email to the candidate and again this is something that of course you will experiment with what what works what does not work because some sometimes too much outreach may not may not be good and then i'll just confirm i'll not confirm because it'll start sending emails to these candidates right now which is not what i intend but that's you you get get the point and it's how easy face on that um surjit that's a really intuitive way i mean even just you could probably people watching this on silent could understand exactly what what was going on when you were putting that outreach sequence together.

44:03So that's well done on the UX and design team on that. That's fantastic. Yeah. Now, the interesting thing here is, look, when the replies come in, Emma will automatically read and tell you who is actually interested, right? And you'll open a list of interested people instead of just like the open inbox of everything. So that's an interesting thing to look at. Now, the other thing here is I can also see I sourced a bunch of new candidates from outside. I can actually configure and push them back into my ATS. So I can say for an account executive job, initial screen, and I can save the changes. So it will actually save this candidate back into my application tracking system as well, Whether I'm using Greenhouse, Ashby, or we connect to all the popular ATSs out there, workday, everything you can imagine, you can just push it back there as well.

45:06So it's two-way sync happening with all your applications. Our goal is not to keep you captive, right? You can go somewhere else and use it, but you can see how simple it is to use directly inside Emma. Okay, so that's our quick demo. I have one more thing to show you, which is, let's say we have got some replies and so forth. And now we need to actually do the next task, which is how do I pre-screen these candidates? And sometimes it's like the same question you are asking again and again, like, do you have work authorization? do have some basic experience, this pre-screening that a recruiter is doing, the first call, first half an hour, meet and greet, and so on.

45:56And recruiters don't have time. So candidates are waiting for sometimes weeks to get actually the pre-screened, scheduled. And we have solved that with our voice AI employees. So I'll quickly show you how you can create a voice AI employee and actually, so a voice employee called the candidate and do the pre-screening, just having a conversation over voice. So if you are ready, I'll go and show you that. It is actually a lot of fun because you can, I usually have mess with it and have some fun. So maybe the demo will break or it's actually a real live product, but usually it handles it very well.

46:41So let's see. Let's see. Credit to you for doing it, Surajit. I appreciate it. But yeah, I think the voice is going to be a game changer. Once this is active and mainstreamed, I think this is the way in which a lot of people are going to choose to respond. Yeah. All right. So let's see. Ask me to stop if you can't hear anything.

47:07Surojit Chatterjee: Hi, Matt. Thanks for taking the call. I'm Cassidy from Acme, Inc., calling about the enterprise account executive role you applied for? This should only take a few minutes. A couple of questions from me, then time for yours. Sound good? Sounds good. Perfect. Let's get started then. I see you speak English, Spanish, and French. Which language would you like to do this in? Let's try Spanish. Claro, cambiamos a español. Dos preguntas rápidas. Estás autorizado para trabajar en Estados Unidos sin patrocínio. ¿Y cuál es tu periodo de preaviso? Actually, can you switch back to English, please? Of course.

47:44Surojit Chatterjee: Are you authorized to work in the U.S. without sponsorship? And what is your notice period? My notice period, I'm not so sure. Where will I find the notice period? It's usually in your employment contract or offer letter. For now, just your best guess is fine. And are you authorized to work in the U.S. without sponsorship? I think my notice period is 30 days. And yes, I think I'm authorized to work, yes. Got it. Now the main one. Tell me about the toughest deal you have closed. So this was a deal with a large provider in a large hospital chain, and they wanted us to build a patient scheduling, patient booking kind of software.

48:26There were a lot of regulatory hurdles, a lot of infosec challenges, and multiple stakeholders. I had to go and convince their head of IT and their head of operations and patient support. It turned out to be a great deal for us. It was multimillion dollars, actually$1.5 million per year for a five-year contract.

48:53Surojit Chatterjee: What was your personal part in getting that over the line? I actually met one of the stakeholders, one of the buyers at an event and pursued him and got a meeting set up and then got a full workshop set up with his team. They were very excited after the workshop and then worked through the navigating the entire stakeholders, managed them all and landed the deal. Of course, I had a whole team working with me and behind me. I won't take all the credit myself. Thanks for sharing this. That's everything on my side. Anything I can help with? Any questions on your mind? Yeah, if you can tell me a little bit about whether this is I can walk from home or this is in person.

49:46Surojit Chatterjee: So it's hybrid. Three days in the New York office, two from home. The office is in a great location. Lunch is on us and we help with the commute. You also get to collaborate with the team in person those days. Does that help? That's great. And it's good that it's not all days in the office. Can you tell me a little bit about the compensation package that I should expect? Comp is a competitive base plus an uncapped commission, roughly half and half. The exact numbers come from the recruiter, but I can pass your question along. Does that help? That's very helpful. Thank you so much. Anything else on your mind?

50:21Surojit Chatterjee: No, that's all for today. When should I expect the next call? Thanks for the time. The team will review and follow up on next steps. Great. Thank you. Bye. Bye. All right. Great interview, Sirajid. I would hire you. You're a very convincing candidate. I don't know. Although I would agree with Michael. I was suddenly expecting you to start busting out some Espanol with us. And I was thinking, what? This guy's got some multi-talented... I actually don't speak Spanish. That's a fake resume. That's great. So we can expect, obviously, so many use cases here. We can expect the sourcing aspect. People are going to be able to pipeline quickly.

51:11You can then set up the outreach call. People are going to take the call. And then, essentially, what happens when the call is completed, What does the recruiter see? They see, I guess, does Emma score the results of the interview, right? Yeah. Emma will automatically score the results of the interview. We'll have the transcript of the call as well for the review. Yeah. And it will feed back in. It will actually suggest whether this person is eligible for the, for example, if I had said I don't have authorization, it will flag that, right, that I don't have work authorization and so on. Yeah. Fantastic.

52:01And just on the next step, is it then handed over to the internal process of the, so at this point, this is the end of the Emma sort of function? No, we don't have enough time to show you everything so it will go into the same application and it will actually you can manage all the stages of the of the pipeline as well as right as as you uh go through but yeah at this point it will uh give the feedback and go into an internal process and and you can the recorders will pick up so yeah i mean i'm excited to see okay how much can can emma do because you're going to absolutely anticipate, okay, initial pre-screen was okay.

52:46Can Emma then eventually maybe go through a more thorough screen and then coordinate maybe a face-to-face interview with the human at some point? All of that, I think, is certainly going to be manifest in the market. That's how it's going to work, right? Absolutely. So what we have done, basically, these are different AI employees. They kind of work together. so recruited employee is sourcing shortlisting and so on then the pre-screened employee takes over channels it back then you'll even have like the whole scheduling etc and onboarding you know after that once let's say you go through the offer process you can have an ai employee that creates the offer based on, of course, your offer templates, et cetera, send the offer, and then onboard the employee.

53:38Let's say employee accepts the offer. So the whole idea is all these manual steps in an employee's lifecycle, the entire hire to retire, where on can we create AI employees to help an organization just get higher velocity and really focus on what's important for the business. See, what's important for the business here is looking at these candidates that AI is pulling up and saying, okay, using your judgment better, right? Instead of figuring out, okay, which data sources will I go to? Where will I find this? How will I send an email? What will I do? You are really focusing on what's important for the business in the end.

54:20Same thing like onboarding an employee, and we had a lot of our customers use us for that use case and of course many more. It's a nightmare in a large organization. You have to create provision someone in like 50 different applications. You know, how do you like L &D training? You have to queue up everything. And it just takes a long time. And same thing if someone changes job, like a job role change, one of our large customers, by the way, that is the first use case we did. And our mind was blown how complex it is in a large enterprise. They have like 30 ,000 to 40 ,000 people every month change location or job role or department.

55:10Something is changing in their job, right? So that means maybe the comp is changing. Maybe the access permissions are changing. Maybe benefits are changing if they move to another state. Payroll may be changing if they move to another country completely, like the payroll provider may change. All of this coordination takes 10 to 12 hours or more for each case. Imagine now, 300 ,000 hours lost every month, millions of hours lost every year. And this is a company, 300 ,000 plus employees worldwide. So it's a massive sap in the productivity where the HR team could have used it a lot better, actually focusing on people and people's growth.

55:57Yeah, and I think this is where we'll look at all of the gaps in previous generation of technology innovation that humans needed to plug. Now we're seeing AI agents plug those gaps and humans are needing to step up into a different function where they add unique value that AI cannot currently do. Very exciting stuff. I mean, I could talk to you for another hour, Surajit, but I'm very conscious of your time and how busy you are. So I don't want to eat any more of your time. But do give us a couple more answers to a couple of questions I have for you. firstly in terms of how people can can kind of I guess do a very quick usage of Emma I mean is it the sourcing and the recruiting aspect that's the cleanest and easiest way to experiment with the product yeah I mean it's the easiest way to get in right of course there are we do a variety of use cases in the HR IT space but you can go to recruiter.mr.ai today and check out the product, test it out.

57:05Actually, we have some free credits already in the account when you create a new account, so you can do a few searches without even paying anything and so on. So give it a try. And I think you can also upgrade to a paid tier if you feel it's working for you or if you want to check it out. Most of our large customers of course we'll have our sales team reach out and help you set up as well so yeah fantastic stuff uh and final question for me sarjit what is down the pipe i mean people who are either currently using emma now or gonna do so or thinking about it but with a 12 month timeline what exciting things are you going to release to those lucky people over the next 12 months Yeah, actually, in the next couple of weeks, we are making a huge, big launch and upgrade of our platform.

58:00But a lot of these workflows that you see are a little bit disparate. We are bringing it together. It's one simple experience, one single glass of pain that any HR admin or IT admin can use and set up very quickly, very simple and self-serve democratizing like complex business processes. Of course, our forward deployed engineers are available and help as well. So that's the big thing that's happening immediately. Going forward, you know, longer term roadmap, we are actually making all of these AI employees really self-improving, recursively self-improving, kind of from human feedback, from every execution they're doing, they're learning and getting better and better.

58:46just like human employees we think about that a lot that similarity between how humans work and how ai work we're doing a lot of r d you know we started the conversation with this news about you can encourage ai to do better better work we're doing a lot of r d on how these ai employees can really be the best collaborators you can have in the future in any enterprise and how can we transform, help transform every enterprise to this new era of AI. Yeah, fantastic stuff. I'm really excited to see where this goes, Surajit. I love the expansive vision. I love the ease of implementation. I love the immediacy of the results.

59:29Like you can literally give it a go now and you can get some outcomes. That would probably cost you like, you know, half a day if you're doing things manually. So go ahead and experiment with it. Surajit Chatterjee, last question for you. What things are you doing now? You're a busy CEO of this rocky ship of a startup. After this call, what's the next thing on your agenda today? A startup CEO's life. I have a bunch of customer calls lined up and I'm talking to the product and engineering team, looking at their roadmap and our OKRs for the quarter, talking to our legal and contracting team from some commercial deals we are doing.

1:00:12partnerships. It's all over the place. It's in a sense, it's everything and anything that comes my way, including sometimes figuring out what to order for food for the team. And so this is it. It looks like Surajit needs to deploy Emma for his own sort of purposes here and get a fleet of AI employees to help you. Surajit Chatterjee, thank you so much for your time. I know you're usually back to back to back to back. So I'm going to let you go now. But thank you so much for spending the time with us, showing us Emma and giving us this wonderful demonstration. I think very exciting to see your progress and I'll be looking forward to see what happens next with this product.

1:00:51Thank you, Sereen. Thank you so much.