TechTalk: Coming to a Consensus: Startup Accelerates Enterprise Innovation with AI Inference via Unanimity

Sri Shenoy: 

Thank you, Fritz. It's a great honor. Looking forward to our conversation. 

Fritz Spencer: 

Likewise. Well, to get us started, I would love if you could give us a brief background on yourself. 

Sri Shenoy: 

Yeah, absolutely. I studied computer science. I've been an engineer for about 20 years. Worked in various companies, Nokia, Motorola, Google, Cisco, held different positions both as an engineer and manager. And throughout the time I was kind of curious about the stuff just outside of my work, whether it's AI, whether it's blockchain. So in 2021, I had this sort of urge and calling to quit my job and start this company, which is Aloha Protocol. 

Fritz Spencer: 

And we spoke briefly before about Aloha Protocol. It's a suite of products. You have a lot of different things that you are developing. Tell us briefly about some of them and maybe some of them that you want to highlight. 

Sri Shenoy: 

Absolutely. So Aloha Protocol is, as you mentioned, it's a suite of products. Some of them are related. SoSo, the primary focus is AI, and there is different products that focus on different areas. SoSo, at its core, Aloha Protocol is built on the principle that not one AI is going to be good at everything like not one foundational AI model when I say an AI. So it's built on the principle that the products that we built are going to use different models, but also there'll be some amount of stuff that is duplicated, meaning certain critical decisions, critical processing is done by multiple AIs, and then we measure if they all come to a consensus. So it's similar to humans when they make important decisions, it's seldom the case that one person makes this highly critical decision in an enterprise, it's usually a board of directors or a committee of some sort. SoSo, we want to take that approach and apply to AI. So that's the sort of the foundational underlying principles behind Aloha. And then we built our primary product today is called Elo and Elo in an enterprise setting, it helps companies understand how their salespeople and their support people are performing and what the customers are saying and where the opportunity to do better is with the help of AI. 

Fritz Spencer: 

Interesting. So it sounds like from the protocol point, you have a jury almost of all the AIs coming together on a consensus, unanimous decisions only, and those are being pushed through as approved responses, and that sounds really interesting. We can dive into that more as well. But tell me more about Elo. I'd love to know what you guys have really built and what you integrated and maybe even who's using it and why. 

Sri Shenoy: 

Absolutely. So if you think about customer support, primarily it's very much human-centric. When you pick up your phone and if you call your bank or your airline or your insurance company or what have you, you expect some person to pick up the phone and help you, or even if it's sales, enterprise sales or any other situation, you have people that typically work in a call center type environment and then speak with their customers. SoSo, some of the things that happened in the last two years, this industry is undergoing a massive change with respect to AI because AI can very effectively listen to these conversations, create transcripts, and understand what is being said. And there are multiple reasons to do it. One is to just capture data, what the customers are saying, it's an automatic way of capturing it and saving it in the CRM and so on. The other one is to sort of give real-time updates to the supervisor to alert them, "Hey, this call isn't going so well, maybe somebody else needs to jump in," and also to the agent. If the customer is asking some question which they don't know the answer to, the AI is already listening and it knows the answer, it knows where to go look up and then present that to the agent so that it seems like the agent knows, has all the answers very quickly. And then there are multiple applications beyond it to analyze, hey, what happened in 10,000 phone calls today? Who's performing better? Who needs training? A multitude of opportunities. And if I may, this is where we begin with, but where we are going towards, some of the features we are building is a step further, which is now that the AI knows all of the conversations, let's say in a sales environment, for example, market sales, AI knows hundreds of thousands of phone calls and conversations and it can go and identify those prospects which has a potential to be a customer but didn't quite become a customer. The AI can analyze those conversations and create follow-ups in terms of emails or voicemails or text messages, andmessages and then try to basically get them to be a customer or if it's already a customer, give them a better support. And then the last piece to that is AI speaking with people. I mean, this is sort of what everybody expects we'll get to, but nobody really has nailed this technology. But there are different versions of it. And we have a version that's sort of fine-tuned for certain verticals where when you pick up the phone, it's an AI that talks to you. You still have the opportunity to talk to a human if you so choose, but there are situations where you may prefer AI and we hope that there will come a time where people will prefer to talk to an AI than a human. 

Fritz Spencer: 

It's very interesting that you mentioned about AI because I know that a lot of people are very apprehensive about it. They find them clunky,clunky; they find them uninformed and sometimes tedious to use. And I'm sure you're going to face some obstacles with that. It sounds like you haven't really had a lot of obstacles developing the technology with your background in so many different technological spaces, but maybe there's some obstacles on the human front that you might face. Can you tell me a little bit about some of the obstacles that you guys are facing and maybe some of your plans for those? 

Sri Shenoy: 

SoSo, I mean, you are absolutely right. I think that technology isn't quite there where it's a better experience for somebody to talk to an AI than a human. Absolutely not. I think everybody wants to talk to a real human because it just sounds and feels more natural. But I think the speed at which the technology is changing, it's going to be a matter of time before the AIAI is going to surpass the humans. And I don't mean to say that people are just going to want to talk to an AI, it kind of sounds impersonal like, "Hey, there's no human touch. It's just a lifeless machine." But I think if you really think about it, it's similar in the sense of a self-driving carcar, right? I mean, when the self-driving cars first came about, if you ask majoritythe majority of people, they would just cringe at that idea. But I think we are much more accepting of it. And I think there will come a time where people think self-driving cars to be safer than humans. So actually, I was listening to the CEO of Waymo recently give an interview where she said, why Waymo is the future, she said, or any self-driving technology, Waymo is creating the best driver in the world, meaning all of the Waymo cars, there may be thousand cars, 10,000 cars being driven right this moment, they're all being driven by one AI and that AI is the best driver in the world. So that's what we are trying to build. So if a company can recreate an AI, which is their best salesperson, and the superpower with that salesperson is that it can talk to thousand people at the same time. And that's one benefit for the company, and it learns from all of their best salespeople and then learns all the techniques of how to serve their customer. But from the customer standpoint, the AI will always know the answer. There will not be a wait time, and it'll always be in the best mood. And I think those are some of the benefits, but I completely agree that we're not there yet, but I think we will be. 

Fritz Spencer: 

Yeah, I agree. I think it's down the road. I don't know how far, but it's certainly in the process. SoSo, I'd like to jump back to Elo as well. I want to talk about some of the obstacles that you might have faced there, or maybe even what exactly is the key problem that Elo is solving for its customers? 

Sri Shenoy: 

The key problems that we are solving, we are that we have some customers in tax debt resolution, and we are entering mortgage services, mortgage sales. SoSo, one of the key problems is supervision, automated AI-based supervision. If I am a company, I have 10,000 phone calls a day, I must have a hundred, 200 agents answering the phones. And there's usually somebody new that requires more training. And as much as a company you offer the training, not everybody is going to learn in the same pace. And you have no idea the people that are representing your company, what they're saying to the customer. Are they representing the company correctly? Are they overpromising if it's a sales environment? And it's often the case that they're overpromised, and then you're not able to deliver. SoSo, there are so many of these problems. And traditionally this was resolved by having an entirely separate team listening to every single phone call. SoSo, there are a hundred people on the phone talking and another a hundred people listening, and this is some of the ways in which this problem was tackled and it wasn't a perfect way to handle it. So primarily that's one of the features that we went to market with and the customers really benefited from. And then along the way, we found so many other little features that had multiple benefits such as capturing all the data, populating their CRM, and just things that are sort of more adjacent to supervising, just overall making the sales process or the support process better. 

Fritz Spencer: 

Very interesting. It sounds highly customizable. It sounds like you guys have done a lot of training on your models, and it sounds like you guys have definitely solved the redundancy in the industry for that specific instance that you gave. I'm curious to know what's next. What are you guys looking to do in the future? 

Sri Shenoy: 

Going beyond Elo, we are building a product in the... it's called MoneyBun, moneybun.com. Our first product is elogpt.com. So what we realized is that Elo specializes in taking calls in a private enterprise call center environment. We received some inbound requests, "Hey, can you do the same for earning calls of public companies?" There are about five, 6,000 public companies in the US alone, and every company publishes an earning call, which is completely public every quarter along with the earning report. They also have to file their quarterly filing with the SEC. SoSo, we are building a product that analyzes all of it and makemakes it much more digestible and easier for the investors to, at any given time, analyze 5,000 company details and make the best investment or trading positions. So that's our next product in the pipeline. And also improve Elo, like I was saying, the AI that talks to people, that's one of the features we are focusing on. And alsoalso, for Elo expand the verticals. What we realized is that the solutions we're building is kind of very vertical specific. If you have to enter a new market such as insurance, sales or mortgage, there's a little bit of re-engineering to customize your product for those industries. So those are things that we are focusing on. 

Fritz Spencer: 

That's super interesting, and I'm so glad that you brought up SEC filings because obviously as an accountant and as an auditor, I live and breathe those every day. So that's super interesting to hear about it. I love it. And I want to end with this last thing. You mentioned in your brief background, your history of working with large Fortune 500 companies, and then you had this moment, you decided, "I don't want to do that anymore." Can you tell me more about either that pivotal moment or another pivotal moment that was very significant to you in your career? 

Sri Shenoy: 

One of the pivotal moments, I would say, just in the history of Aloha, I was a first time entrepreneur. I had learned, read a lot, but it's never the same when you are in the trenches and doing it because almost all the companiescompanies, I worked for were bigger companies, and it's a totally new... everything you think you know, you get humbled. And I went through that experience. It felt very cold because when you don't have a recognizable brand attached to your name and you try to make a sale and you get rejection after rejection, after rejection, because here's some person from a company that I've never heard of. So it's really kind of getting over that learning curve. And alsoalso, another one, another important pivotal moment is really always, always don't build a product in vacuum. I always felt, "Oh, so you come up with an idea, you build something and then you try to sell." And that rarely works unless you're a super genius like Steve Jobs or somebody. What we discovered then, that's how Elo came about, is that we go to a customercustomer, and we tell them, "Hey, we are going to build this for you for free for six months or a year in exchange for feedback and then data." Just have the opportunity for them to tell us what they want free of cost. And then that's really... again, I mean that's what worked for us to embed ourselves in a setting like that. And then that really made all the difference. So yeah, I think that's what led us to the success that we are now experiencing. 

Fritz Spencer: 

Wow. I'm so glad to hear it, and it's such an interesting story that you've shared with us, Sri. Thank you so much for taking the time to join us on the podcast. It's been an absolute pleasure. 

Sri Shenoy: 

Pleasure is all mine. Thanks. Thank you, Fritz, and it's been a great honor speaking with you. 

Fritz Spencer: 

And a special thanks to our listeners for tuning into Tech Talk, the entrepreneurs and innovators who turn to EisnerAmper for accounting, tax, and advisory solutions to help propel their success. Subscribe to EisnerAmper's podcast to listen to more Tech Talk episodes or visit eisneramper.com for more tech news that you can use. 

^{Transcribed by Rev.com}