Saving Customers & Agents Time with AI | Puneet Mehta from Netomi

DISCLAIMER: Below is an AI generated transcript. There could be a few typos but it should be at least 90% accurate. Watch video or listen to the podcast for the full experience!

Alexander Ferguson 0:00
In this episode, we interview Puneet Mehta, founder and CEO of Netomi. They’re a machine intelligence company that uses AI to turn customer service into a competitive edge. Putting spent much of his career as a tech entrepreneur, as well as on Wall Street building trading AI. His current venture started in 2015. He is focused on customer service and what he calls a workplace multiplier for enterprise companies striving to give back time to customers and agents, we discussed met his ambitious plans to become a publicly traded company over the next few years. Thank you so much money for joining me. And I’m excited to learn more about Netomi. And kind of where it started, where you’re going that market segment that you are helping more about the technology behind it, and how you, as a entrepreneur and CEO are innovating and staying ahead of the curve. So to start us off, what year did you start your this business that you’re with now?

Puneet Mehta 0:58
So thanks for having me. I started this company in 2015 2015.

Alexander Ferguson 1:04
And the industry that you’re primarily focusing on serving is what?

Puneet Mehta 1:09
So we are an AI platform for customer service. Our AI works right alongside human customer service teams, and it is a workforce multiplier.

Alexander Ferguson 1:21
A workforce multiplier, I like the concept multiplier is a great idea at scaling. So being able to help primarily enterprise companies, is that your target segment?

Puneet Mehta 1:30
Yes. So we mostly work with, with companies that have a sizable customer service team. And that is interested in kind of having AI take off some of the things that can we can from their plate so they can focus on human employees can focus on kind of the bigger picture, and the kind of work that that they can really learn from and keep doing on long term.

Alexander Ferguson 1:55
Got it? Obviously, with enterprise, there are a lot of customer service increase for them. So being able to effectively scale and solve customer service issues is very valuable. And that’s the pain point that you’re focused on. Is that correct?

Puneet Mehta 2:11
Yeah, I mean, if we look at the state of customer service today, I think both sides are challenged, the companies are challenged, because this just is such an intensive task to have every query be answered by a person, right. And the consumers are feeling challenged, because they want instant resolution in today’s world of instant gratification where you can get anything from, you know, the best macaroons in San Francisco to, you know, an iPhone delivered at your doorstep in 40 minutes, you want your issues to be resolved instantly. So I think both sides are feeling challenged. And that’s kind of where we come in and provide that workforce multiplier. You know, where companies can do this without spending a lot of money, they can make their agent teams happier, because the agents don’t want to do the repeatable tasks, and the consumers are happier. And that’s why we’ve seen for all of our customers, the customer satisfaction goes up. Our idea is to give time back, you know, we’re giving time back to agents, we’re giving time back to consumers. And everything that we do with AI is focused on the idea of getting time back.

Alexander Ferguson 3:23
I like that you look at it from both sides, both the enterprise helping them scale that their workforce, but also the consumer, I had an experience recently, where I tweeted a large enterprise wanting to solve a problem I had. And it was two days later, I got a tweet back. And I was annoyed that it took two days. And if an AI agent had replied at least acknowledge that it was received, I probably would have been a much happier customer. But we’ve also had experiences where we’ve been on automated telephone calls, and they’re not listening and understanding. So it’s it’s the balance of what it can do. And but your platform allows to roll over where human can take over. Is that correct? Yeah.

Puneet Mehta 4:01
Yeah, yeah. So So you know, there are two things we do every time we deploy AI, we figure out, you know, the capability part, but we also figure out the authority part. You know, you want to train your AI when it comes to certain things really, really well, potentially better than what people can do on those things. Not everything, there’s certain things that people do better. And we really believe in AI teaming up with people and not replacing people. But the second aspect is the you know, the quality piece, even if AI is capable. Do you want to give it authority to do all the things or do you want to be selective in the beginning? So so we are the only platform in the world that looks at it with these two moving parts. That’s why our customers really love us because you know, it makes it almost zero risk to launch an AI within a large organization,

Alexander Ferguson 4:52
giving the control to the enterprise of how much control itself the AI agent has. How many customers clients do you have right now that are using this platform?

Puneet Mehta 5:04
I cannot share an, you know, an exact number because we are still a privately held company. We haven’t gone public yet. But, you know, it’s a, it’s a large, healthy base of, you know, some fortune 500 companies that are using it already.

Alexander Ferguson 5:20
I imagined for each enterprise, though the quantity of customers you end up serving because the amount of customers each enterprise has, it scales very rapidly of how many end consumers you’re actually supporting. Like, right. So for then the technology that makes you difference, that’s the engine that runs the whole thing. Tell me more about your intellectual property, the, the AI, the machine learning behind it.

Puneet Mehta 5:43
Yeah, absolutely. So, you know, we have built our platform of a technique called reinforcement learning. So, you know, with the emergence of deep learning, which you see in everything from Tesla’s autopilot to what was behind AlphaGo, the AI that beat the Go champion of the world, you know, we are bringing that similar level of technology to customer service. So, you know, in the, maybe even five, seven years ago, this type of tech was reserved for, you know, NASA or Wall Street or, you know, places where it was, it was having significant impact, but had not made to customer service yet. So we are really bringing that kind of tech to customer service. So we filed a patent last year on how you can have any, I can have a full stateful conversation, and have both long term and short term memory for an Kelley context forward. So all the things that you would you would want, you don’t want your customer service agents to replicate what happens in 54 states, you know, but unfortunately, that’s the state of customer service, every time you pick up the phone, they’re like, you have to remind them again, who you are, and what problem you’re facing. Sometimes when you bounce from agent to agent, you have to repeat the whole thing again. And it’s just such an incredibly, you know, a wasteful use of our time, we are both for the agent and the company and the consumer, it’s costing us both time and peace of mind and money. And that’s kind of where we come in. And that’s why we want to make sure we build something that works for both sides. But yeah, the core tech is deep reinforcement learning. It’s a neural network that learns in three ways. You know, learns, by instruction learns, by example, and learns by experience. So it’s very close to kind of how people learn

Alexander Ferguson 7:34
this reinforcement learning, is it based off of each new is an instance like for new enterprise? Is it learning that enterprise? Or is it a collective learning across all of your customers?

Puneet Mehta 7:44
So it’s both, um, you know, obviously, our customers data is of utmost importance to us. You know, we are a security first AI company, where we don’t want one customer’s data and learning making another customer to protect kind of their privacy and their business. But at the same time, there are certain things which we can abstract and bring out layer above. And that’s the common shared learning across the neural network. Also, we invest in training, keep training this neural network, not just with our customers data, but with your general interaction data that’s available on the open Internet, and, you know, we crowdsource a lot of training data, we also have synthetic data from where we are able to coach the AI based on what a successful, you know, human interaction could look like. So it’s a little bit of both where we have this platform that goes across customers, which is constantly learning and evolving and growing. So the newer customers are coming on board are getting more advanced platform by the day. But then some of the learnings from a business stay within their computer to

Alexander Ferguson 8:51
garden, what kind of partnerships do you currently have with other companies or with other platforms that really enables and extends your own product?

Puneet Mehta 9:00
So we partner with, you know, agentless companies, obviously, that’s an obvious one. Companies like Zendesk, we integrate with Salesforce, or they’re both CRM part in the service cloud. Yeah, so, we are we have a few integrations in that particular area, we also integrate with some systems of record, you know, like Shopify is of the word and demand where Magento and we are able to support kind of closed loop transaction capability, right within a conversation with that, and then you know, we have built a platform and an open API gateway. So we are able to integrate with any system out there, whether it is you know, shipping system or tracking system and ERP engine. So our customers are essentially taking these business processes, finding the conversation part of it, bringing it into the, into the AI, but then at the end of the day, every conversation results in some outcome. So we able to go ultimately go integrate with the systems of record where you can also influence

Alexander Ferguson 10:06
is the core interaction of an end consumer would be some sort of chat conversation like a chatbot type interface.

Puneet Mehta 10:14
It could be chat, it could be email, it could be, you know, chat between an app chat on the bed chat with a social channel. So all of those,

Alexander Ferguson 10:24
you cover all the major different ways a consumer could interact, such as like messenger, or Twitter, or what are the different platforms that you integrate

Puneet Mehta 10:34
all of the above anywhere where a customer can chat with the brand, we support that. We’re also doing some early work on The Voice side. Something I can probably talk about next year, and it’s not that launches. But we also started doing some work on voice.

Alexander Ferguson 10:53
What’s one difficulty that you’ve had to face in the last year or two, as you’ve been progressing that another entrepreneur can learn from, as you’ve had to get to the point where you are now?

Puneet Mehta 11:03
Yeah, um, you know, I think I can talk about the one difficulty that AI faces or industry, which we face, and I think most of the AI companies face, you know, We specially face in this particular space, because the hype around this was, or has been quite high. And then there are companies out there, you know, that might pretend to be a companies and then go sell to a large enterprise. And the large enterprise would be like, oh, you know, this isn’t quite what was promised. And then I think that that ruins it for the telco system. So we’ve seen, you know, companies that would go out there and would have the right marketing and messaging but not the right. And I think that that just is a disservice to the entire AI community.

Alexander Ferguson 11:55
So where do you see your company in five years from now?

Puneet Mehta 11:59
And five years, I think we should be a publicly traded company. Wow. So yeah, I think, you know, considering the size of this opportunity, and the market pool that we are expecting in the next couple of years, I think we should be able to grow quite rapidly. It was truly a foundation year for us, and 2020, we should hit the ground, running on the good market side and go start acquiring customers.

Alexander Ferguson 12:24
While I’m excited to then see that progression as you grow and become that publicly traded company, you’ve already mentioned one hurdle that you’re gonna have to get around of those, the bad taste is what’s in some people’s mouths. But do you see any other hurdles that you’re going to need to overcome or the industry itself to, to or for you specifically, to realize that vision? Yeah,

Puneet Mehta 12:44
I think all the other factors are in our control, that’s the only factor that’s not control. You know, all the other factors, whether it’s managing risk for these large companies as introduced AI in their organization, it’s baked into the product, you know, like, we don’t expect them to operate on full autonomous mode from day one, hey, I can just trap responses for you from day one. So. So there are things whether it is on the deployment side, you know how to shrink the time to try time to deploy time to value, like all those things we are actively working on. So they’re all in our control. And I think we have, we have the mall in a pretty good place, where we can be a really good partner for large enterprises looking to bring in AI into their customer service team. I think only thing outside our control is if these large companies go and end up buying from a company that doesn’t truly have AI and they cannot tell good from bad, they will have a bad experience. And you know, we cannot cannot help them if they pick the wrong car.

Alexander Ferguson 13:43
For you personally, as you move forward, how are you innovating? How are you continue to learn? Where do you go for news? To keep you fresh? Yeah,

Puneet Mehta 13:53
um, you know, that’s an interesting question, especially when, when you’re creating something that does not exist. You know, a lot of times you, you have to kind of interpret market signals a little differently than how they’re interpreted in the media, right? You just have to kind of ask the FBI, you might have to ask multiple buyers, right? You, if a large company makes a certain decision, you’re like, Hey, why did this happen? So you have to go beyond the obvious. And, you know, but but there are a lot of inspiring podcasts these days. So more than just, I would say, staying on top of what’s happening in industry because we’re creating a new category. So hopefully, we are the ones breaking the news. News rather than, you know, consuming breaking news, but I think in general better when it comes to younger companies and, you know, creating categories. There are a few podcasts like, you know, a lot of good books being written now, the knowledge on this has gotten a lot more organized.

Alexander Ferguson 14:59
specifically on that point, is there is there one, one book or podcast or audio book that you would recommend someone else should read? Yeah, I mean, in

Puneet Mehta 15:09
case of startups, I would really recommend entrepreneurs read all Graham’s essays, Paul Graham, who started by Combinator, you know, I, I’ve read some of his essays multiple times, and, you know, in different times if they have different meaning, so, so I totally recommend that there’s a really good podcast on how I built this. You know, I, I’m fond of that, and, you know, they keep having new episodes. You know, on the book side, you know, the hard thing about hard things. I think that’s what it’s called, the one that Ben Horowitz wrote from Andreessen Horowitz. So that’s one of my favorites. Zero to One is quite awesome beauty. So

Alexander Ferguson 16:02
that’s just a few that we put on our reading list. I love it. That’s, that’s powerful. Thank you so much for joining us there. Where would you recommend that people go next, to learn more, and kind of start their journey learning more about you, and to connect?

Puneet Mehta 16:19
Yeah, my website is the best place in the Tomi and e to emi.com. And we have a really good resources section, where there are a lot of videos, where if you want to learn about AI, specifically about customer service, we have no resources there. We keep publishing a lot of blog posts. Some of these are industry specific, even, you know, some for retail for travel, hospitality. So, so watch out for that. And then we also have a good newsletter, where we can condense all of this information and send out so yeah, if anybody wants to learn about AI about customer service near our resources section on it for me.

Alexander Ferguson 16:58
That concludes the audio version of this episode. To see the original and more visit our UpTech Report YouTube channel. If you know a tech company, we should interview you can nominate them at UpTech report.com. Or if you just prefer to listen, make sure you’re subscribed to this series on Apple podcasts, Spotify or your favorite podcasting app