On this episode of the Six Five On The Road at Zoholics 2024, hosts Cory Johnson and Lisa Martin are joined by Zoho Corporation‘s Ram Ramamoorthy, Director of AI Research, for a conversation on how Zoho is addressing the evolving AI challenges within the enterprise landscape.
Their discussion covers:
- Zoho’s holistic approach to integrating AI across its suite of products
- The most significant AI challenges enterprises face today and Zoho’s solutions
- Future trends in AI technology Zoho is investing in
- The role of data privacy and security in Zoho’s AI strategy
- Real-world applications and success stories of AI in Zoho’s ecosystem
Learn more at Zoho Corporation.
Watch the video below, and be sure to subscribe to our YouTube channel, so you never miss an episode.
Or listen to the audio 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.
Transcript:
Cory Johnson: We are back at Zooholics in Austin, Texas, Six Five On The Road. Is it really on the road if Six Five is usually in Austin, Texas?
Lisa Martin: That’s true.
Cory Johnson: All right. It’s more on the road. We’re at the Austin Convention Center. We’re at the Zooholics Convention. A fascinating company that a lot of people are just learning about, a lot of people have been using for a long time. And as in all things technology right now, the big story is AI.
Lisa Martin: It is AI and we’re so happy to have the Head… Sorry, the Director of Research at Zoho Labs. Ram is here from Zoho. Thanks so much for joining us. Lots of news today going on, but the AI strategy, it just became the last 18 months rocket ship with GenAI.
Cory Johnson: Certainly they like you. Now they care.
Lisa Martin: But everybody needs it, but they don’t know, how do we start? Talk to us about some of the things that you’re doing there?
Ram Ramamoorthy: Well, enterprise AI is a very different ballgame. We have seen the consumer world take off with AI. You get consumer-based large language models that can help you write your essays, that can help you plan a vacation. But what does it have to do with enterprise? I’ve always seen enterprise technology as a late mover, and it takes the advantage of the late movement. We see the AI moment coming to the enterprise right now, but not without its own set of challenges. Limited amounts of data, very sensitive information that can be your competitive advantage. Now, how do we make sure we get the highest quality of AI predictions with the limited amount of data that we have?
Lisa Martin: Sure.
Ram Ramamoorthy: And that is where Zoho specializes in. We’ve always been a privacy and a security first company. We apply the same principles to AI, get the maximum accuracy with the limited amount of data that you have.
Cory Johnson: Well, and I want to talk about the limited amount of data. The limited amount of data you have is because it’s customer data and the customers are not as big as say the internet, which is what the large language models most of us in the country are trained on.
Ram Ramamoorthy: Absolutely.
Cory Johnson: So it’s almost like you’re using these small language model.
Ram Ramamoorthy: Yeah, that’s right. We use the small language models. In fact, we call it contextual intelligence, where we contextually combine models-
Cory Johnson: I’ll license the phrase small language model to you. I make you a deal.
Lisa Martin: I think I heard that before.
Ram Ramamoorthy: So the way we work is one, we have been having a lot of narrow AI models, models that can do only one thing at a time. Say for instance, a sentiment analysis model for customer support emails cannot do sentiment analysis on customer support chat.
Cory Johnson: Interesting.
Ram Ramamoorthy: So one model that does only one thing. Now, when these LLMs became big, what came is models that generalize very well, models on which you can plan a vacation on, models on which you can build AI chat and so on. But in the enterprise, you really don’t have that much of data, but you need higher levels of accuracy. So put these smaller models, narrow models, wherever possible. But I’m not brushing away the advantages of large language models because they have this ability to generalize very well and they have an emergent capability, which these narrow models don’t have.
So we want our customers to get advantage of both the worlds, in fact. So use narrow models wherever possible. And when you need the emergent behavior, bring in a small model if the data warrants. If it needs to be more emergent-
Cory Johnson: But that also requires you to know what you’re doing… I don’t even know what you’re doing like there’s fools operating at the gears. But I think a lot of what is happening with AI at this point in 2024 is people experimenting with what it can do. So a narrow language, a narrow application of AI is harder because it’s narrow and you don’t really know what you’re not doing yet.
Ram Ramamoorthy: Yeah, and the challenge is, how do you embed that into your everyday enterprise process? There is a workflow that goes through multiple departments that needs approvals, and where do you put in AI so that it comes in at the right place? In a nutshell, I would say we’ve been talking about process automation for a decade now. AI is decision making. Can you automate the decisions for you backed by your historical data? And AI is just statistics powered by historical information. You know what is normal at a given point in time? Thanks to all the learning that these AI models can do. So successfully imbibing that in your everyday business workflows is going to be really helpful.
Lisa Martin: One of the things I heard you say this morning in the media and analyst session was that Zoho’s… You talked about its commitment to principles that matter. You just mentioned one of them, context, contextual, truthful, privacy focused, value-driven. How do these tenets, if you will, guide Zoho in delivering AI that really truly benefits the 100,000,000 users you have now?
Ram Ramamoorthy: Yeah. So let me walk you through an example. Now, let’s say we have been using optical character recognition model to automate your expenses that you submit on a business trip to your finance team. So now we have-
Cory Johnson: Or reads the receipt.
Ram Ramamoorthy: Yeah. So receipt, click a picture, it extracts the information and sends it up.
Cory Johnson: Right.
Ram Ramamoorthy: So this has been around for four or five years now.
Cory Johnson: Sure.
Ram Ramamoorthy: These are narrow models where-
Cory Johnson: Concur, ramp that you guys go. Sure.
Ram Ramamoorthy: Yeah. So you do that. Now, what happens next is hundreds of receipts end up in an admin and it’s his or her job to actually look through it and approve it or go back, ask a question to the employee on why this was done? Now there is a lot of context here. Now, let’s say I had the marketing for a particular region, and let’s say I booked this convention center with my corporate card and the system knows that, “Oh, Ramprakash, when he goes to this region, he puts all these bigger expenses on this card.” Now, let’s say I go to a different region and it’s just my food and stay in flights. And each organization has a different travel policy. Let’s say hotel rooms cannot exceed beyond $200 a night or whatever.
So there are so many contexts where you’re hyper-personalizing this expense automation based on my role, based on my location, based on my team, and at the same time, based on the event that I’m traveling to. So put all these together, that makes it a very contextual AI thing. And ultimately, your mean time taken to approve an expense report comes down by a notch.
Lisa Martin: Oh, dramatically.
Ram Ramamoorthy: Dramatically.
Lisa Martin: Yes. Yes. Well, what you’re describing from a context perspective is so important. And you talk about hyper-personalization, and we expect that in our consumer lives that we can do any transaction, whether it’s a ride-share or e-commerce. And the experience is going to be relevant-
Ram Ramamoorthy: Absolutely
Lisa Martin: Hyper-personalized in a secure, private way, but relevant to us. And so that consumerization has bled over into our business lives that we want the same, we demand. It’s not a want, it’s a demand.
Ram Ramamoorthy: It’s a demand.
Lisa Martin: From customers across every industry. I imagine that there’s not one industry where that demand is going to be there.
Ram Ramamoorthy: Yeah, and we see that. And with enterprises, the data becomes very sensitive. Now, let’s say your expense report, your OCR engine that processes your expenses is subsidized by leaking all the data. And I can see that a lot of your company employees are spending with Starbucks. So maybe it’s an opportunity for another coffee shop to go run a marketing campaign. It should not work that way. Right? It seems crazy because your personal data, I don’t pay for my search engine, I don’t pay for my social network, but I give back a lot of my personal information. But the challenge is-
Cory Johnson: That’s how you pay.
Ram Ramamoorthy: Yeah, that’s how you pay.
Lisa Martin: Exactly. It’s currency.
Ram Ramamoorthy: And the problem is, see, my smartphone knows that I’m broke. My smartphone knows that maybe I’m borderline diabetic, but it does not stop it from showing fancy milkshake ads at 11:00 PM in the night, pushing me toward, or it keeps showing ads to upgrade my phone to the latest one. So you see the difference. Your data is your data and it should be working in your benefit.
Lisa Martin: Yes.
Ram Ramamoorthy: Maybe it should push me to exercise more. Maybe it should push me to, “Okay, it’s okay, your phone is still good. Maybe don’t go for that upgrade.” But it doesn’t work that way.
Lisa Martin: No.
Ram Ramamoorthy: Now, the same thing cannot be applied to the enterprise because that’s your competitive advantage. And we at Zoho, we’ve been very particular on our privacy stance, not just with AI, with everything we do. Right from the days we started, we didn’t have ads even in our free plans, and that’s how we work. So we apply the same tenets to AI and that is what makes us so unique in this feat.
Cory Johnson: I think another thing that’s unique is, and this is my perspective as a public markets, US stock market person, but the fact that you’re a private company and your approach towards AI, at least what you say your approach to AI is is, “We’re not trying to put AI in front of our customers. We’re just trying to get stuff that works better in front of our customers.” I’m paraphrasing, and I’m thinking about the massive criticism Google has gotten for not having its own ChatGPT and not having it… Which it has now I guess, but for not having an AI strategy and AI. What do customers want? And you guys just can skip the, “We have an AI strategy,” part of things to just go right to what the customers need.
Ram Ramamoorthy: Yeah.
Cory Johnson: And I think part of it’s because you’re a private company.
Ram Ramamoorthy: Yeah. We do a lot of unorthodox things. In fact, the decision to host on our own data centers wasn’t well received when we did that. “Why don’t you go with an AWS, or why don’t you go with an Azure?” But when you play the long game, the decision to build offices in tier two, tier three cities didn’t really start off very well. But today, here we are, most of our workforce is in tier two, tier three cities stopping urban-
Cory Johnson: You guys have been in Austin for a long time.
Lisa Martin: And now McAllen.
Ram Ramamoorthy: And now what? McAllen?
Lisa Martin: Yeah.
Ram Ramamoorthy: McAllen is our biggest growing office in the US. Our CO himself lives in a remote south Indian village called Tenkasi.
Lisa Martin: Yes.
Ram Ramamoorthy: We’re like four hours away from the nearest international airport.
Cory Johnson: Right.
Ram Ramamoorthy: So we are a very, very unique company. We believe in playing the long game, staying private has really helped and we can go fail with our experience. So 2011, I started the first AI research as intern, and any big company would’ve possibly gone and acquired somebody to do it or gotten in a bunch of PhDs to get started to do it.
Cory Johnson: Right.
Ram Ramamoorthy: But I was just a fresh programmer out of college and I got the opportunity, and here we are. So a lot of things might sound crazy or unique when we start, but I’ve seen over the years, the conviction has really paid off at the end.
Lisa Martin: And that’s one of the things that Shridhar talked about this morning was that conviction. We’ve heard it from every guest we’ve had on the program today, customers, Raju was even talking about it. And it’s one of those things that I always like to understand the culture of a company. You hear about it, but Corey and I have been talking about this too, there’s a uniqueness here that isn’t just unique in saying so. It’s authentic.
And I think that we’ve had that message reinforced. I’m sure it’s reinforced in the partner community, but in the last few minutes that we have here, walk us through, you’ve been there a long time. The evolution of the company where that philosophy to stay private, to grow organically has been maintained as the company has grown?
Ram Ramamoorthy: So it is a motivation, right? It is building that technical know-how, and taking your time, taking your time to get things right, understand how a deep… You work 10 months on AI, you’re an AI expert. Right? So give that engineer the time and space to get things up to speed. So we run a lot of unorthodox projects. So for example, we are investing in things like generally GPUs… People talk about AI, people talk about video processing, gaming and all that, but we use a lot of GPUs for our database acceleration workloads. And we find that it’s relatively greener to throw in GPUs than hundreds of CPUs on databases.
Cory Johnson: Right.
Ram Ramamoorthy: So very, very unique. And these experiments might fail. A public company might not be able to do it, but we have the liberty. We start small, we fail fast, and that has really helped. And we fail fast and only when it scales… See, 2011, I was the first intern who started AI and it was only in 2015 I got my second colleague. So taking things slow, failing fast has really helped.
Lisa Martin: That’s awesome. Ram, thank you for joining us on the program and walking us through what you’re doing from a research perspective in labs, how you’re helping customers embrace it on a pretty quick pace, doing it securely, doing it privately, and the democratization of access. We appreciate your insights and your time. This has been a great conversation.
Ram Ramamoorthy: Pleasure is mine. Really appreciate the opportunity. Enjoyed it.
Cory Johnson: We’re going to have one last look at what’s happening here at Zooholics, with some of the best analysts from Six Five and from Futurum Group, to really get a sense of why this matters. And we’re going to have that for you right here from Austin, Texas. Six Five On The Road, right after this.
Author Information
Cory Johnson is the Futurum’s Chief Market Strategist and the host of the Drill Down podcast.
His peripatetic career has seen him in prominent roles as a hedge fund portfolio manager and investor, technology journalist and broadcaster. Fundamentally he’s an entrepreneur -- helping to start media companies such as TheStreet.com, the Industry Standard, Slam (the world’s best-selling basketball magazine) and Vibe. He was CNBC’s first Silicon Valley correspondent and later helped create the TV show Bloomberg West for Bloomberg TV and the radio show and podcast Bloomberg Advantage. He was a senior executive at the blockchain startup Ripple, a portfolio manager for Kingsford Capital and a principal at the Forensic Research Group.
Johnson is also an advisor to Braintrust, C3.ai, Prolly AI, Provenance Bio, Stringr and serves as a delegate to the Episcopal Diocese of California.
Lisa Martin is a technology correspondent and former NASA scientist who has made a significant impact in the tech industry. After earning a masters in cell and molecular biology, she worked on high-profile NASA projects that flew in space before further exploring her artistic side as a tech storyteller. As a respected marketer and broadcaster, she's interviewed industry giants and thought leaders like Michael Dell, Pat Gelsinger, Suze Orman and Deepak Chopra, as she has a talent for making complex technical concepts accessible to both insiders and laypeople. With her unique blend of science, marketing, and broadcasting experience, Lisa provides insightful analysis on the latest tech trends and innovations. Today, she's a prominent figure in the tech media landscape, appearing on platforms like "The Watch List" and iHeartRadio, sharing her expertise and passion for science and technology with a wide audience.
