On this episode of The Six Five – On The Road Partner Edition, hosts Daniel Newman and Patrick Moorhead welcome Ada’s Mike Gozzo, Chief Product & Technology Officer, for a discussion during MongoDB’s .local NYC event.
Their discussion covers:
- Mike shares Ada’s founding story and its mission
- We discuss the ever-present topic of Generative AI
- How the use of MongoDB’s platform came about in their organization and how Ada’s team is using the developer platform, within the FinTech market
- Mike also shares his advice for other C-level executives looking to incorporate generative AI into their strategy
You can watch the full video here:
You can listen to the conversation here:
Disclaimer: The Six Five webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.
Patrick Moorhead: Hi, this is Pat Moorhead, and the Six Five is live here in New York City at MongoDB.local. As you can hear from all the great sound around us, we’re on the show floor. There’s a lot of stuff going on, big product announcements. We’re talking to MongoDB executives, we’re talking to customers, we’re talking to partners, and it has been great so far, Daniel.
Daniel Newman: Yeah, we’ve kind of covered it all here and it’s been a really great day. From the developer to the C-suite, we are seeing what it’s going to take to really usher the generative AI era into that foundational data compute network and infrastructure that really every company right now is up against. And one of the best ways, I think, Pat, here on the Six Five to really add value to our audience is to talk to those that are building.
Patrick Moorhead: That’s right … to actually using it.
Daniel Newman: It’s great to talk to the vendors, and that’s a lot of where our time is spent. But a lot of where our knowledge is gained is those partners and those customers that are putting the tech to use that actually validate those moments of punditry in which we enjoy so much.
Patrick Moorhead: That’s right.
Daniel Newman: And I think we should do that again.
Patrick Moorhead: Absolutely. Mike, how are you doing? Thanks for coming on the show.
Mike Gozzo: Thanks.
Patrick Moorhead: So you run products at Ada, and by the way, I was really impressed at what you showed because it wasn’t like hypotheticals. It was like, here is what we’re doing. And anyways, congratulations on what you did. You educated a lot of people on the possibilities of what could be done. Thanks for coming on the show.
Mike Gozzo: My pleasure. Thanks for having me.
Daniel Newman: Yeah, it’s a lot of fun. And by the way, I heard the story that you were on the stage sharing your story and you hadn’t actually been pre-briefed about all the announcements. So there’s a possibility this-
Mike Gozzo: That all came together really quickly. Yeah.
Daniel Newman: There’s a possibility that your story probably could have landed even better because you could have started to tie some of these new Search, the Vector, some of the stream and started telling how this is going to help. But now that you’ve got some time to think about it, we can use this as a platform to talk about that. Hold on, before you do that. I want to get the backstory because it’s not necessarily a household name, but I walked away from your presentation being exciting, wanting to learn more, tell our audience, tell us about the company ‘s founding story.
Mike Gozzo: Sure thing. So, Ada was founded in 2016, in Toronto, Canada. Most of the team is Canadian, but we’re all over the place and pretty remote. We started pretty much with the one singular focus of automating the most customer service inquiries across all channels, across all languages, across all modalities with the least amount of effort. And since 2016, we’ve come a long way. I mean, we’ve raised over 191 million US in funding and grown quite a bit. We’re over 300 employees, but more importantly, we’re over 300 companies. And many of these are upper mid-market enterprise tier customers, names you would recognize like Meta, Verizon, AT&T, and so many others. And we sit behind the scenes automating their customer service interactions. There’s a good chance that people that are viewing us now in the Six Five have already used Ada and not realized that they were speaking to a bot at all.
Patrick Moorhead: It is amazing how much time as industry analysts we’ve been spending on generative AI and typically, or at the front of these things because people want to be … they need to be educated and there’s value. And enterprise is asking, Hey, what type of approach should we spend on this? It struck me in your presentation that’s just the space you’re in, I just know generative AI can absolutely improve the customer experiences that you’re providing them.
So my question is, what has changed over the past few months for you? I mean, deluged with requests, I mean the excitement must be pretty high.
Mike Gozzo: It’s been validation. I think as an industry, we definitely have gone through different hype cycles. 2016 when the company was founded was the height of one bot cycle. We all knew it. We felt it, but it was underwhelming. You might remember Mark Zuckerberg going on stage and announcing a weather bot, raised millions in funding and then kind of fizzled. And there was a bot backlash, that was real and that actually happened. But generative AI has changed that because now that we have the transform …
We’ve had the transformer for years, we’ve been working for it for years. But as these models get larger, and as RLHF, human feedback, reinforcement learning through human feedback has come onto the scene, the quality of the understanding, reasoning and generation of the technology has started to equal and in some cases surpass human’s ability. And that’s the real unlock. Right now the question is not can we automate so many customer service requests that we completely change the dynamics of a contact center? We’re pretty certain we can do that. Everybody in the industry and aid is leading on this, is trying to answer how do we do that in a safe, accurate, and reliable way, not “how do we do it?” It’s a very different problem.
Daniel Newman: I mean I think every boardroom right now is kind of sprinkling down on people in your role like, “Get it out.”
Mike Gozzo: Exactly.
Daniel Newman: I mean, I’ve heard about CEO departures and just having conflicts because of the pace, meaning there’s those of … in your case that you’re kind of saying yes, but we have to do it with these constraints. And then there’s others saying, get it out, get it out, get it out. And I think we’re going to find … This is where we hear about companies having to stop their employees from taking sensitive data and dumping it into these platforms because people don’t fully understand what’s at stake. And it’s not just even their mistake, it’s the PII data. It’s what went into these things that people didn’t think about. There’s a massive opportunity for growth, but there’s also a massive opportunity for people to get this wrong.
Mike Gozzo: Put it this way, if I was a traditional agent desktop vendor and I was just taken private or acquired by some other company, I would be looking at my investment committee and my partners around the table if I was a P-firm and asking, holy hell, did we overpay for this in a world where now generative AI has demonstrated its promise, and I think that disruption is hitting customer service in a big way because the application is natural, but it’s really going to hit so many other processes because in this wave of technology, the true power of generative AI is not in automating digital processes, which is what software’s done for the last four decades.
The true power of it is in automating human processes. And this is new for all of us in the industry. We don’t know how enterprise SaaS, if there even will be a thing called enterprise SaaS in the future, will change with this technology yet.
Daniel Newman: I actually think it’ll be different. I’ve got my prediction of that. We’ll hold that for later. That’s where me and I get to do our pundit thing. Mike, talk a little bit about the relationship … You’re a fast-growing startup. It’s 190 million funding raised, hundreds of customers, everybody’s used your stuff, the decision to partner, build, depend on MongoDB. What drove that?
Mike Gozzo: I mean, really in the earliest days of the company, we were thinking about Velociti first. And the key reason why we went to MongoDB at all was because it was a NoSQL database and the stack was extremely clean for us to use. We knew that the company was going to change over time. We started in customer service, we stayed in customer service, but we could have gone so many different ways. And having the flexibility and the ability to just pivot on a dime was really important. And we saw that as we advanced the company, brought in new channels, new modalities, one data store that could be easily extended without crazy migrations that would really support our needs was absolutely cure for Mongo. And look, there’s been challengers along the route. We’ve had people build even API-compatible versions of Mongos from some cloud vendors, but we’ve always stayed the path with Atlas and Mongo because the performance is there, the support from the team is great, and we believe in having less dependency on one central cloud vendor that MongoDB allows us to have.
Patrick Moorhead: So both MongoDB and Ada can deliver horizontal value. But if I look at the history of IT, the biggest value is when you can go vertical and in a more customized way, serve the needs of your customers. And they love it and they’re willing to pay more. And that’s goodness for everybody.
One vertical area that MongoDB announced today was the financial services vertical. And I also noticed from your website, that seems to be a high value vertical for you. And I’m curious, what incremental value are the two of you bringing together to the market? How are you using MongoDB in this space?
Mike Gozzo: One of the most interesting things that we’ve been involved with on the MongoDB side is a concept called queryable encryption. And that allows us to have an end-to-end representation of the conversation that’s encrypted. Why it matters is that contrary to popular belief, the way you work with a generative AI system is yes, you could throw massive amounts of data at it and train it, but it’s actually more interesting to distill the actual message and guidance you want to give the system, particularly when you’re working with large-scale commercial models like GPT-4. So in these modes, Ada doesn’t need to actually read the customer’s conversation. We just need the bot manager’s coaching and feedback on the quality of conversation to bring through to the system. So the more that we can secure, the better it is, particularly for fintech and financial services companies.
Daniel Newman: It’s interesting. He should talk to the CEO of One AI. I don’t know if you had a chance here. We had him on the show. We’ll have to send you the link because he’s doing everything about language. And he was talking on the show here about the ability to take all the data that’s been created in say a meeting, and then actually the ability to really find those high value generativity and then give actionable … because it’s like there’s so much data being created, it’s almost how do you take out of it rather than putting in and generating. And he talked about how you can generate, but it’s not really right and wrong, it’s not black and white, but finding the value in the generated data is harder to extract. So it was an interesting conversation we had here.
Mike Gozzo: So when you think of generative AI is like that … I think that when we first saw ChatGPT as the world, the natural inclination was to use it the way that we always have with Google and ask it a question. And we think it’s this Oracle that’s all-knowing. But the real way to think about generative AI is that it’s understanding the relationships between words in a very deep, non-intuitive way. And because of that, it can reason. So you don’t have to think of these systems in terms of how do I get them to answer a question? You have to shift your mental model to be, how do I use them as an assistant to resolve a problem where you’re actually asking it to break down problems into steps, refine those steps, clarify them and help you think. And that’s what our bot is doing behind the scenes. We’re not just asking a question and answering, we’re asking the bot to reason through a multi-step problem that will be different every time you ask that question. And we’ll generate the solution by following its steps.
Patrick Moorhead: Mike, so the generative AI conversation has gone to all levels out there. I mean from my family and to the dinner table, but also up to the boardroom because the evaluations of companies, quite frankly, lately are being determined of, hey, what’s your … It used to be like, what’s your web strategy, what’s your e-commerce strategy, what’s your generative AI strategy. And we’re seeing market caps move significantly. So you’re obviously in a position to give advice to C-Suite. And I’d like to know what advice are you giving to the C-suite right now on generative AI and these capabilities?
Mike Gozzo: That’s a really good question. The biggest thing that we could say is … and this is a little self-serving. You need to lean into this and realize that your competitors are moving on this already. This is a blockbuster Netflix moment that’s hitting everybody, Ada included because we’re a seven-year-old company and it took us time and we had a lead up on it. But it took us time to modernize our tech stack to be able to lead into this revolution. But every single business in every industry is now faced with an attack vector on it that didn’t exist before. So we’re really urging companies to imagine that at this moment, they’re blockbuster. And what advice would they give to the blockbuster CEO when Netflix walked in and offered to sell them the business? You should really think through what that will say.
Daniel Newman: I think it was 50 million by the way.
Mike Gozzo: Yeah, right.
Patrick Moorhead: Well, it’s funny we’ve seen different waves and we call this Web 1.0, but it was travel agents, it was content, it was stockbrokers. But for the most part, there were a lot of elements of industry that were really untouched. And a lot of the C-suites that I talked to were looking at different areas, and were looking at hardcore customer support. And yeah, six years ago, right, the chatbot was going to revolutionize that space. Well, it took maybe a couple steps forward, but we ended up with the human handoff, which I view as a technological failure.
But looking at the call center, looking at accounts payable, accounts receivable, a lot of areas that weren’t really that impacted by that Web 1.0. But I think there’s a lot of opportunity to increase productivity and come up with a lot happier customers.
Mike Gozzo: I think that that’s exactly right. As I mentioned, reasoning is the core thing that is being replaced by generative AI and large language models that was untouched before, and that’s really where we’re going to shine and where the opportunity for disruption is in the industry.
Daniel Newman: There’s a lot more to talk about here. We’re going to leave it here at this time. But Mike, I look forward to following Ada. Congratulations on all the success, and thanks for joining us here on the Six Five.
Mike Gozzo: My pleasure. Thank you very much.
Daniel Newman: All right everyone, thanks for tuning in here. This was a really interesting one. We have all seen the pressures of getting and building your next generation of technology around generative AI. This is one that you probably wanted to tune into. If you missed it, rewind back, you can hear it all. If you did like what you saw, hit that subscribe button. Join us for all of the episodes here on the Six Five, but we’re on the road here at MongoDB.local in New York. But for Patrick, for myself, time to say goodbye. Tune in to all the episodes. See you later.