The following interview is a conversation we had with Ofer Ronen, CEO & Co-Founder of Tomato.ai, on our podcast Category Visionaries. You can view the full episode here: $12 Million Raised to Build the Accent Softening Category
Brett
Welcome to Category Visionaries, the show dedicated to exploring exciting visions for the future from the founders who are on the front lines building it. In each episode we’ll speak with a visionary Founder who’s building a new category or reimagining an existing one. We’ll learn about the problem they solve, how their technology works, and unpack their vision for the future. I’m your host, Brett Stapper, CEO of Front Lines Media. Now let’s dive right into today’s episode. Hey everyone and welcome back to Category Visionaries. Today we’re speaking with Oprah Ronin, CEO and Co-Founder of Tomato AI, an accent softening platform that’s raised 12 million in funding. Ofer, how are you?
Ofer Ronen
I’m doing well. Hi Brett, thanks for having me here.
Brett
Yeah, no problem. Super excited to have you here. So as I was preparing for this interview, I saw that this is now your third contact center solution company. Did you just grow up just dreaming of, you know, data or contact centers? Like where did that come from? Why are you on round three now?
Ofer Ronen
Yeah, doesn’t everyone love contact centers?
Brett
Come on, astronaut or a contact center?
Ofer Ronen
Well, I’ll tell you like AI is cool. Everyone would agree with that. And AI these days there’s so much you can do with language. And it turns out call centers have a lot of language to process, whether in real time or after, let’s say a call or chat. And so it’s a good marriage of technology and like a problem space. And so I have juicy problems to solve and so I’m on my third contact center AI solution or startup. But really it’s because it’s a large market and AI can be used across different use cases. So yeah, I figured I’ll keep going, you know, while I’ve had a good luck with it.
Brett
And for the first two, can you just explain a bit more about what problem you’d set out to solve with those first two?
Ofer Ronen
Yeah. So I was at Google inside of their incubator called area120 and the problem I was working on was how do you build virtual agents that can field complex conversations, multi turn complex conversations. Often what we’re used to when we call, let’s say Verizon to say maybe one word and we’re careful to say it well, billing and hope that it catches it. But what we did at Verizon and elsewhere through the work at Google was to be able to just speak naturally and have follow up questions and you know, all the more complex ways we interact with humans can be done now with Virtual agents. So ultimately what we did there with the first one called Chat Base was take millions of past conversations between humans and use that to model more robust virtual agents for voice and chat.
Ofer Ronen
And we did it with some of the largest companies and Verizon was one of them. It’s public information, but yeah, I structured and closed over half a billion dollars in call center ideals for large companies like that.
Brett
Wow. What’s the key to getting a deal like that done?
Ofer Ronen
Yeah, you know, that’s maybe a few hours of conversation, but if I was to summarize, it’s patience because there’s a lot of like gatekeepers and people you have to like convince lots of pilots and even bake offs, like we had to compete against other entrenched vendors to unseat them. It’s about an 18 month process. We did three a hundred million plus deals and yeah, and each one is like roughly 18 months start to finish. Obviously worth it if you can close it, but it’s a pretty extended process.
Brett
So that was startup one, what about startup two?
Ofer Ronen
Yeah, so then I had so much fun doing that one. We sold it to Google Cloud. It was an internal transaction at Google that was like our exit. And a year after selling it to Google Cloud, I went back to the incubator area 120 and the second contact center AI startup there was related to. Well, what we did was like when you call customer support, sometimes they don’t solve your problem the first time. In fact, it’s very common that they don’t. And so then you call back and it’s very expensive, all these callbacks. And they tried to reduce those calls back. Instead we created a solution that proactively with AI identifies, oh, there’s no resolution to this call. What if we reached out and tried to solve it using automated means to avoid a callback, which was expensive.
Ofer Ronen
Callbacks can be as high as like nine bucks a call if a US agent responds. And so if you could send a text and maybe get them engaged in some automated process to resolve the issue they’ve raised, that’s still open now it’s saving money. So that was the idea. And yeah, with that one, I would say it required like more effort to get to work, more customization per company. I still think it’s a good idea, but unfortunately like in the end area, 120 closed big companies, you know, spin up divisions, closed down divisions. So that idea I think still has legs, hasn’t been realized yet.
Brett
Well, perfect segue to talk about number three. So you’re Back at it again like we talked about. What problem are you setting out to solve this time? Is it the same problem as startup 2 or a totally different problem?
Ofer Ronen
So first of all, with this third one, I actually ended up leaving Google after seven years there and decide to go back to entrepreneurship outside of Google because that’s what I was doing before I sold Google some startup, then did these couple of internal startups. So I went back to doing my own startup. I teamed up with James, my CTO. He was also there working on Contact center AI leading speech to text. Text to speech. Okay, so what’s the new idea this time we decide to help humans do better with AI instead of trying to automate way humans. So we switch teams to what I call human superpowers with AI.
Ofer Ronen
So what we’re doing is when an agent talks to a customer and let’s say they have a heavy accent, often the customer judges them as less capable and is struggles to understand what they’re saying. And so with AI we can do something called accent softening. As the agent speaks, their herd is having less accent and that helps everyone involved. Like the agent doesn’t have to repeat themselves as much. They suffer less abuse. The customer can relax and more easily get through the call because it’s a more familiar way of pronouncing words that they are hearing. So that’s the idea. Very challenging problem. We’re doing real time speech AI. As someone speaks, we’re modifying their voice to sound. In this case with less accent. That’s the new one.
Brett
Do you just give everyone a British accent? I feel like whenever I speak with someone with a UK accent, I just trust them more.
Ofer Ronen
Good insight. So yeah, so there’s accents that we look up to and accents we look down on. And all you need to do is get an accent that they look up to and respect. And you’re right, Americans respect British accents. That’s not a bad idea. Currently we’re giving folks Ohio neutral American accents. But yeah, in the future we’ll have British accents as well. But again, as long as you’re getting them to an accent that they look up to, that’s what the key changes that you want to go for.
Brett
Ohio is a good choice as well. I feel like the Midwest. I feel like you respect people in the Midwest. Good. The hard working values. Yeah, work then is it happening in real time?
Ofer Ronen
Yeah. So we have a speech to speech AI model. What that means is it doesn’t transcribe it and generate the speech, it just transforms the speech in the speech Realm. And yeah, it happens in a split second. As you speak, we have different versions. Some of them you can’t even tell even with camera on that it’s happening because it’s so low latency. Other ones, there’s a slight delay, but you can still have like a back and forth conversation. So depending on. It’s like the difference between quality versus latency. So, you know, depending on the use case, you might pick one version or the other.
Brett
I know you had touched on that there’s, you know, kind of different paths. You’re on the side of empowering the humans. A lot of the other conversations I’ve had, a lot of the stuff I read online, it’s all about, let’s get rid of humans. We don’t need humans. It’s going to be all, you know, AI.
Brett
So talk to us about that decision, because that sounds like a pretty interesting and also a very intentional strategic decision.
Ofer Ronen
Yeah. So I think AI that’s going to help automate and AI that will help humans and both will play a role in the next few years. And we just saw, having been in the call center space, my Co-Founder and I saw this opportunity to help human agents. That’s not there. And there’s a whole suite of solutions. And, you know, the beachhead we thought is accent AI, but you could do grammar AI. Translation, AI Enthusiasm. AI. You can raise the enthusiasm. There’s a lot you can do.
Ofer Ronen
Now, we think that if you can at a lower price point, where, you know, if you think about some of these offshore markets where it’s cheaper to hire people, if you can make them sound better for, let’s say, US customers, then you can really open up that market, create all these jobs for people there and still have the human touch. So a lot of people that call, they want to talk to someone. They want to negotiate, they want to have some compassion. They want to. They don’t know how to describe their issue. There’s all these issues that come up. They just want to talk to someone. Their Internet’s out they want to talk to someone. So we can make that possible still, but at lower price point, but higher quality experience. That’s call centers.
Ofer Ronen
Now if we look at beyond that, there’s meetings like this where people could use this as well. The creator economy could do it. If I’m streaming and I’m Scottish, what if I sounded American or Indian or Filipino, Whoever’s listening could hear me the way they’re used to. So there’s lots of like offshoot kind of markets, but we’re starting with call centers.
Brett
And how do you think about that balance? You know, given you have technology that can be used in like a thousand different use cases, how do you know when to decide it’s time to expand to another use case or go deeper and try to get more market share in contact centers?
Ofer Ronen
So I’ve previously sold two startups to Google1000.com as CEO and one of the things I’ve learned is focus is so important. And so you want to do one thing really well and be known for it. And then once you’ve done that, then you think about adding, expanding your focus. So we’re focused on Filipino and Indian accents. I would throw in there Pakistani as well, because that’s similar to Indian and that’s our focus and we’re going to do that really well to start. And once we do that well, we’ll keep expanding like that focus. The mistake that I think startups often make is they go too wide too soon and then they’re not known for anything and they don’t do anything exceptionally well. So that’s how we think about that.
Brett
What about from a marketing perspective? How would you summarize the current marketing strategy? And has that evolved at all since you were founded a few years ago?
Ofer Ronen
Yeah, so I’ve always done B2B startups and that have been like sales focused, but obviously you still want to do brand building and be found easily. Easily and be known. And so yeah, there’s been investment in SEO that’s been ongoing, a little bit of like work to get covered in the press, but we’ll do more of that and ads. You know, as we are ready to grow with customers, we’re getting like a very healthy inbound stream of interest that’s good for this stage we’re in. And then yeah, the idea is like, well, hit the gas more as we are ready for more business. I would also say there’ll be a whole like sales go to market motion that we’ll invest in at that point.
Brett
Now let’s Say that I’m running a contact center. Say I have a thousand people out india. And I’m like, okay, this accent problem is, you know, is hurting us. Is accent softening, Is that a category? Like, when I go out and search those phrases and like, is that what I’m searching for? Or, like, what am I even searching for or considering to solve this problem?
Ofer Ronen
Yeah. So because it’s a nascent category, there’s a few terms people use, and over time there’ll probably be some consolidation. But what I would say is some of these names are prescriptive about features they think are important. So whether you call it accent matching or accent softening, you know, that hints as to what you’re going for with your solution. But a neutral name would be accent AI. You know, that’s a more neutral name. The term, unfortunately, that’s used currently is accent neutralization. Like, that’s what a lot of call centers call it, accent neutralization. And they have trainers come and help people neutralize their accent. I think that’s a bit of a harsh term like neutralizing. And so that’s why we chose to go with accent softening, which I think more people will sign up for. We’re reducing the accent hurt. We’re softening it.
Ofer Ronen
We’re not replacing the accent. Okay, but to answer your question, there’s a few terms out there and still no clear winner.
Brett
Yeah, I think softening sounds a bit better than what was the other term?
Ofer Ronen
Neutralization.
Brett
Yeah, this sounds a bit aggressive, I would say. So that makes a lot of sense. How big is that market for, you know, companies coming in and trying to neutralize your contact center’s accents?
Ofer Ronen
Yeah, so this is one of those cases where there’ll be a market forming around the accent AI kind of solution in the coming years and winning vendors will emerge. And obviously we’re positioning ourselves to win in that market. It’s definitely like a multibillion dollar industry in that if you take, for example, the back of the envelope math is, take the 15 million agents around the world, assume that they all will benefit in one way or another from this kind of tech times. Maybe 50, $60 per seat. You know, that gives you some order of magnitude for a total addressable market or total potential market. But obviously not all of it will be addressable right away. So they’ll be in phases.
Brett
Given you’ve spent so much time in this space, what’s a surprising fact about contact centers? That people like me who their own experience is calling centers. Like, what’s something surprising or interesting about the business that would surprise us.
Ofer Ronen
So, yeah, there’s a lot of adoption of AI that’s starting to happen. And the way that shows up is the agents will be more aware of the context as you start speaking to them maybe than ever. And a lot of manual work in the call center will be automated. Like after a call, they spend all this time to capture notes, you know, so there’s some of these things are really. With the large language, models are starting to be automated well, and then solutions like agent assist are helping feed answers to the agent so they sound more knowledgeable in the moment. But really AI is helping them do that. So, yeah, the goal of AI in the call center is to be seamless and like disappear or not be something the customer notices.
Ofer Ronen
Even the work we’re doing, our goal is to, like, help the agent sound better, but not somehow make the interaction sound awkward or anything like that. So this is what I would say, like, there’s a lot of call centers now experimenting and adopting AI, and that’s going to make interactions less frustrating, faster, more pleasing to the customers.
Brett
Who are you selling to? Who’s like the typical ICP that you’re selling to?
Ofer Ronen
So there’s often at call centers, someone that’s thinking about adopting new solutions and experimenting with new solutions. So I would say that’s often the executive that would first want to lean in here. They have an experimental budget and they’re going to run pilots with new solutions, new categories of solutions, and pick their winning vendor.
Brett
Is that like digital transformation or like a Chief Digital officer? Titles like that?
Ofer Ronen
Yeah. Yeah. So in some companies, that was the title. You know, the titles are so different across companies that sometimes hard to, like, say exactly what that is. So it’s more just who’s that person running pilots and trying new toys and sometimes it’s the QA team. Like, I’ve seen that be the team sometimes, yeah, the Chief Digital Officer. Sometimes it’s like Chief Architect. Like, that’s the title of those people. I’ve seen all kinds of titles.
Brett
You know, it seems like for a lot of industries, AI is somewhat new. But if I look at this space, I know there’s been a lot of different solutions talked about for a long Time I mentioned I had a PR firm or half a PR firm, I worked with a company in 2016 that was bringing AI to contact center. So it’s not totally new like I think some other industries where it’s one or two years in. What’s the general sentiment in the industry around AI?
Brett
Like, have a lot of these people been burned by AI solutions that were too early, so now they’re a bit skeptical when someone like you comes in? Or like, what’s the general sentiment around AI?
Ofer Ronen
Well, one lens by which you can look at that is there’s a new study that I got like a preview of coming out of Gartner, you know, the analyst company, and they were asking so like 2/3 of their customers outsource, use outsourcers and often the call centers are outsourced. And so it used to be that the reason to outsource was cost savings. That was like the number one reason. What’s happening now is the top two reasons are automation and gen AI. Like, those are their top two reasons to outsource. So what’s that speaks to is these large, like Fortune 500s that often, like use outsourcers for call centers are no longer just thinking, oh, it’s a way of saving money, but instead they see it as a way of like getting ahead and differentiating with their call center.
Ofer Ronen
And that’s top of mind for them. So, no, AI is not losing its kind of shine yet. There’s still like a ton of interest to figure out. What can we do with it? How do we get the most out of it?
Brett
What do you make of companies like Klarna? I don’t know if you’ve been following what they’re doing in the news, but you know, they pissed off everyone in SaaS probably when I said, we don’t need Salesforce, we don’t need work day. Like three or four months ago, they probably pissed off people like you who said, you know, we don’t need contact centers anymore. I think it was 700 employees they were able to reduce. Like, is that just because it’s like very specific problems that their customers had? So like, the humans, like just weren’t needed or how are they able to do that?
Ofer Ronen
You know, a lot of people are saying that if this was real, they should have like published more details. First of all, the question is, like, what was really done there and did they characterize it correctly? Now I generally think companies should focus on their core strengths and product and not be wasting time rebuilding CRM and other solutions. Just because I think it can be a distraction and you’re not going to get all the features and capabilities in any reasonable time that you might need from these third party solutions that are in court. You’re a business, so I know. So if we’re talking about like yeah, using AI to generate your SaaS software, I’m still skeptical as to whether that’s a good use of your time.
Ofer Ronen
I do think over time maybe there’ll be like templates out there and maybe lower price points for these products. But I worry about like again, it goes back to focus. Remember we said you gotta focus and this can be a distraction. Now if you’re talking about automating the call center, this is where you have to differentiate between chat and voice. Chat is a lot easier to automate versus voice. In the voice channel, support and sales people are more verbose. There’s emotions. It’s a 10x more complex to field that and automate it fully. Klarna, likely because they’re a tech company, has most of their tickets in the chat space where both the types of people I use chat, they’re looking to as quickly as possible without a lot of small talk, get in and out. And so anyway there’s more of an opportunity to automate.
Ofer Ronen
But if we’re talking about the voice channel and that’s where we operate, you know, a huge percent of the interactions go to humans today just because of the extra challenges.
Brett
I know you had mentioned Gartner there earlier. What are your views when it comes to analyst relations? Is it critical to go to market? Is it nice to have in the go to market?
Ofer Ronen
What’s the current view? Yeah, so I believe if you’re doing enterprise sales, you need to have the analysts cover you in order to get better traction and shorten sales cycles. So yeah, it’s a shortcut for getting a stamp of approval. That helps a lot with enterprise sales. If you’re doing SMB sales, you don’t need that. Or consumer sales, I don’t think you need it.
Brett
What about fundraising? So as I mentioned there in the intro, over 12 million raised to date. What have you learned about fundraising throughout this journey?
Ofer Ronen
Yeah, okay, so one thing I would say is you either have traction or track record when you’re fundraising. I’ve sold three startups and so I leaned into my track record to raise the money needed to get this company off the ground. We’re building AI models from the ground up, foundational models. And so there’s more of a cost involved in that. R&D kind of cost. And so I would say if you’re lucky to have a track record, use that. But that’s the minority of entrepreneurs. If you don’t. Yeah, then it’s about traction. And what you need is to get the right traction for, let’s say, the stage you’re in. And, you know, we can talk about what that might look like, but that’s the game you’re playing. And so those are the two camps.
Ofer Ronen
Other than that, I would say have amazing demos to win the heart and then all the information they need to win their minds, you need both. So invest a proper amount in the demo because a lot of people don’t do that well enough.
Brett
As you reflect on this journey, what would you say has been the most important go to market decision you’ve made so far?
Ofer Ronen
I would say to invest early in SEO because that takes time to marinate and so you want to do it before you need it to be good way before. So early investment.
Brett
In that final question for you, let’s zoom out three to five years into the future. What’s the big picture vision look like?
Ofer Ronen
Okay, so the category we’re in is real time speech to speech AI. And I believe that category will become significant. If you think about all the conversations humans are having, if that can be enhanced, that’s big. And so we’re starting with accent AI, but then there’s grammar AI and translation AI and enthusiasm AI, and all these things you could do to make someone sound better to the listener, improve KPIs. So I just feel like that’s going to be a significant category that will emerge and we’re looking to be a leader in that space.
Brett
How do you think about the ROI or like quantifying the ROI of an Ohio accent, for example?
Ofer Ronen
Yeah. So I think more about how do we quantify the ROI of a sales call or a support call. And for that I don’t need to reinvent the wheel for support. They have things like average handle, time reduction, callback reduction, CSAP, you know, customer satisfaction improvement. And so I just rely on their KPIs for sales. It’s often revenue or close rates, things like that. So what you’re doing is making sure that those improve. And yeah, you’re going to try the Ohio accent maybe versus the British accent to see what works better.
Brett
Keep me updated. I’m curious about that one. British or Midwest accent.
Ofer Ronen
Curious. I’ll keep you posted.
Brett
That’s good. Amazing. Well, we are up on time, so we’re going to wrap here before we do for anyone that’s listening in that wants to follow along with you. Where should we send them? Where should they go?
Ofer Ronen
Yeah, so you can go tomato.ai where you can request a live demo. And, you know, that’s, I think, a good way to do it. Also, you can reach out on LinkedIn, my full name Ofra Ronan, and connect. And happy to always chat with folks who are interested in talking.
Brett
Amazing. Well, thanks so much for taking the time.
Ofer Ronen
Yeah, thanks so much, Fred.
Brett
This episode of Category Visionaries is brought to you by Front Lines Media, Silicon Valley’s leading podcast production studio. If you’re a B2B Founder looking for help launching and growing your own podcast, visit frontlines.io Podcast. And for the latest episode, search for Category Visionaries on your podcast platform of choice. Thanks for listening and we’ll catch you on the next episode.
Twitter
facebook
linkedin