Episode 259: Simplifying AI for Global Customer Impact with Srinivasan Raghavan

AI has acted as a catalyst for the product managers to become more product builders. What I mean by a product builder is historically or traditionally, the product has been that, look, we'll figure out what the customer needs are, we'll talk to customers, and then we'll get user experience people to build a prototype or a design. We'll initially start with some user research, then we'll have an architecture discussion, then engineering will start building things, then we'll start testing things, right? What AI has allowed is to collapse this whole thing. And I strongly think we will have a lot of product builders going forward. Product management will morph into a product builder where you can do user research, design, build the product specifications, and the design. Stay close to the customers, understand their world deeply, and connect everything that we do, whether it's strategy, roadmap, design, or AI, and backed by the real outcomes that we deliver for them. And the tools and friends will change. Like today, we have a stigma, make tomorrow it could be something else, be lovable, or cloud code, all these things have means to an end. But the mindset of being outcomes-driven and being customer-obsessed will always be relevant. Creating great products isn't just about product managers and their day-to-day interactions with developers. It's about how an organization supports products as a whole, the systems, the processes, and cultures in place that help companies deliver value to their customers. With the help of some boundary-pushing guests and inspiration from your most pressing product questions, we'll dive into this system from every angle and help you think like a great product leader. This is the Product Thinking Podcast. Here's your host, Melissa Perry. Hello, and welcome to another episode of the Product Thinking Podcast. Today, I'm excited to have Srini Raghavan with us. He's a Chief Product Officer at Freshworks, driving the growth and innovation of AI-powered customer and employee experience solutions. Their work impacts 75,000 companies worldwide, including big names like American Express, Disney, Databricks, and the Los Angeles Dodgers. I'm thrilled to dive into his insights on how the industry is changing and explore his strategies for enhancing customer success. Welcome, Srini, it's great to have you. Great, Melissa, thanks for having me. I'm looking forward to the conversation. Me too. I think we've got a lot to dive into on AI as well. So I'm curious, what led you to becoming the Chief Product Officer at Freshworks? Actually, an accidental product manager, believe it or not. As a software engineer, I actually started out as a support software engineer when I was carrying a pager, and when the systems are down, that's when they paged me and I would attend to those problems. So that actually gave me a deep appreciation for the impact that the products have on the customers. And I would wake up in the middle of the night and fix their problems. And later on, I became an engineer, so I would design the products. Gave me deep, and then I went into other functions. The main thing that this taught me is how to build great products and not just to design them on a slide. And over time, I found myself more and more drawn to the why behind what we are building, the customer problem, the business outcome. And that is what pulled me into product management. I had a pretty diverse journey before I got into product management. I worked in finance, I worked in banking, I worked in product strategy and corporate development in different companies, which gave me a 360 degree view of how products are built, especially in a B2B SaaS company, are built with the close collaboration and coordination between different functions in a company. And that, for example, at Five9, I started in corporate strategy, then I went into product management. I led product management and user experience there. And we focused heavily on AI and automation. So we built a platform, this worked for automation and agent assist, which were essentially to make the users in the contact center way more effective, which is the human agents and the leaders. I also ran side in corporate development there. I led three acquisitions, which taught me a lot about how to build versus buy and how to make those decisions on what we build versus where we partner, versus the companies that we could potentially buy. Then at RingCentral, as a chief product officer, I had a chance to work on a broader communication collaboration suite, again, with AI and automation as the key teams. And that experience of operating at scale across multiple product lines and segments was a big stepping stone. So I would say in the last 10 years that I've been a product manager, the team has been building a lot of the AI and automation products, which started my journey at Five9, then at RingCentral and now, this is what naturally led me to be at Freshworks. I'll tell you more as we go along. At Freshworks, we have a generational opportunity for us. I am a chief product officer here. I lead the strategy, product strategy, user experience and operations across the two major product lines that we have, customer experience and employee experience. What really attracted me to come here was the opportunity to bring AI powered enterprise grade experiences to a wide range of companies. You mentioned some of those customers at the introduction. We have 75,000 customers in 120 plus countries, but we do it, we build our products in a way that's accessible, simple. We call it the uncomplicated way of providing experiences to our customers. Wow, that is a lot of customers too. And that's a really big reach across 120 different countries. When you're thinking about servicing that many customers as well, what do you try to instill in your product strategy and philosophy to make sure that it's consistent and that you can serve so many different types of people? Yeah, I think the core of our philosophy is serving the end user. You have 75,000 customers. We have a ton of human agents that are using us, 2 million plus agents or so that are using us and a bunch of leaders and millions of end users are using our products. So we're a B2B SaaS company, which means our products are used by our end customers, which ultimately they use to serve their end users. So the philosophy is the number one thing that we look at when we are designing anything is how do we deliver seamless and uncomplicated experience to the end users and to the customers that are using our products. And that is the number one thing that we consider and simplicity user experience and not just of using the products, but experience of the products, using them on an ongoing basis, how they can take advantage of the ongoing innovations that we get to our customers on a monthly basis. That's what we look at. So it's the user experience and usability is at the core of everything that we do. So in serving a lot of those customers, you talked about how you want to make a great experience, especially for the end user. AI is changing a lot of the ways that we do that, delivering value to our customers, also the way that we do our jobs as product managers. Can you talk a little bit about how you're thinking about incorporating AI at Freshworks? Yeah, I'll talk about it in terms of how we are leveraging AI for building products. I strongly believe that what AI has acted as a catalyst for is for the product managers to become more product builders. What I mean by a product builder is historically or traditionally the product manager has been that, look, we'll figure out what the customer needs are, we'll talk to customers, and then we'll get user experience people to build a prototype or a design. We'll initially start with some user research, which is the research team will do. Then we'll have an architectural discussion. Then engineering will start building things. Then we'll start testing things. What AI has allowed is to collapse this whole thing. And the product manager to be more of a product builder. And I strongly think we will have a lot of product builders going forward. Product management will morph into a product builder where you can do user research, design, build the product specifications and the design. And finally, you can even, some of them could even do an initial prototype and test it with customer before you take it to engineering to actually build a production-grade product. Ready to level up as a product manager? Product Institute offers online courses that teach you the strategy, tools, and data skills you need to make confident product decisions. Access expert-led videos, practical worksheets, and proven frameworks, all at your own pace. This holiday season, get 40% off all courses with code HOLIDAY through December 31st. Visit productinstitute.com and start thinking like a great product manager. Training teams? Email info at productinstitute.com for group options. Just as a fun example, this morning I presented our 2026 roadmap to our CEO and the design team. And every time we do this in a bunch of slides and some applications, et cetera, then I thought, okay, why not do a fun little project? So I used Gemini to build an app that can make our roadmap into a living and breathing application that can be accessed by all the stakeholders across the company. So it was a fun little project for me. So I went from being a product manager to a product builder because literally I was a single person that took the inputs, created the app, and published it to RC Suite to look at what the roadmap is. And that was a fun little project. I think these kinds of projects are evolving from fun little projects to becoming actual production gate. We use Figma Make for design prototyping. We use Cursor for initial prototyping of things that we are building. So there's a lot of tools that the product and the engineering and the user experience team are using. Some of these are still in early stages, right? And some are in higher levels of adoption. Some are not there in terms of adoption. So what we're thinking through is how do we create like a formal training program for the entire product team to evolve from being a product manager to being more of a product builder. So just to add to this, there's actually customer value that you get. Like, you know, you sit in front of a customer and customer says, hey, can you do this? And two days later, you come back to them and say, hey, is this what you wanted? This is an amazing thing, right? Like previously, it used to take like months and months. Now you can do this in like a day or two. And that's an amazing outcome for customers. Yeah, I love the rapid speed that we could turn this stuff around and we could test it. One of the questions I was getting last week was speaking at some product conferences in different companies. And a lot of the leaders were asking me, especially about like rapid prototyping or now that we have cloud code, how you see the role of product manager changing, but also how you see the surrounding roles. So how do you see the roles of like UX evolving? How product managers will work with developers in the future? Where do you plug them in and how do you think it's going to change from what we do now? I think every role is going to evolve. Like we have these discussions in our own team where user experience, user research, product managers, everybody is moving much more faster with the AI tools that we are having and be more efficient. So for example, for early stage product prototyping, we don't necessarily need to have the user experience. Designers don't necessarily actually need to have any PMs. If it's a front-end heavy product where the user experience is the key, they can build a prototype in Figma and they can directly work with the engineering team to get that to production. So like I said, every role, these terminologies that we have today, which is user experience, user research, user experience designer, product manager, all these roles are starting to blend. I think in the future, a really good, it'll all become instead of a really good product manager or a user experience designer, good person that's called a product builder. And this product builder will have the skills of user research, user experience design, building a product and working to get the product to production. Some of them, some of the more advanced ones will actually even build a product and ship it directly. Yeah. I think it's so funny when in New York City and around like 2010, I remember we always had these roles called product designers at some of the smaller companies. And the product designer was like a hybrid product manager and UX designer role. And as we kept going, I saw them disappear out there, but that's how I started, was like a product manager and UX designer in one. And I was decent at the UX, I got better at it over time, but the product management is where it thrives. And as these companies got bigger and bigger, I feel like the role split out more and more. But in the smaller companies, I found it particularly great for product designers because you could take it all the way through, like from product management, all the way through user experience if you had enough of those skills. But I just never saw them hired as much over time. It's like, they very much formally separated. It was like, no one could do UX, you don't do product. It's like, there is no hybrid role. You have to stay out. And I always thought that was very strange. I completely agree. I think that's how typically it has been in B2B SaaS and somewhat in B2C companies as well. But what AI has done, it's acting as a catalyst to bring those things. The nimbleness that you mentioned that smaller companies have, that nimbleness is going to come to the larger companies as well. Some of the larger companies, and a lot of companies are like a conglomerate of smaller products, like smaller teams that build different products. So I think that nimbleness is going to come back and AI is going to act as a catalyst for that to happen. I consider myself to be much more of a product manager than a UX designer, but now I'm getting better and better at UX design. And these tools that are there have me and the rest of the product team in becoming a lot more effective at it. And vice versa on user experience design as well. I see UX designers coming up with amazing ideas on how we can improve the products and what does this value is that can deliver to our customers. And they do amazing job of doing user research. So these roles are starting to blend and we will see this more and more happen going forward. Yeah, and I think it's going to be important as well for what you were just saying, like product managers to learn a lot about the user experience parts too. I've met a lot of product managers who are not good at user experience design. And that's the part that makes me fearful about the rapid prototyping landscape because I think Figma Make and all of those things level, they're great at the UI components, but then if you don't actually tell it how to do the experience, they don't string them together very well. It could be very disjointed. And I worry about product managers advocating all of that just to that prototyping piece and saying, oh, they'll figure out the UX. But I like how you were talking about it where you're getting better and you're learning and you're stepping through it. That's right. We actually use a bunch of tools. We do live research with our customers and end users. And when we say customers, there's a person that sort of buys your product and then there's a person that uses your product. Typically those two are not the same in a B2B context because you have a procurement person or a head of IT or a head of support that buys the product, but then the agents and the end users are using your product. So we actually talk directly to our end users. We do user research by talking to our end users. And those conversations happen across the board. We have 200 plus product managers and user experience, people that talk to our end customers, but all the data is synthesized and we have a tool. Those conversations are recorded, they're synthesized, and we get those insights into building the actual product when we are in that ideation phase. So it's a loop. Look, AI can help in doing the design, but the humans is what make it more effective by bringing those insights from the conversations from real users into making this design way more effective. I love that. I love that you're ensuring that actual real insights are getting in there as well. When you think about the landscape of AI, there's also the components of how do we think about how it improves the value for our customers. We talked about speed as one of those aspects too, but what other things have you been looking at using AI for to help your customers achieve their value? Yeah, at Freshworks, let me talk about how we're helping our customers at Freshworks across the customer experience and the employee experiences. We think of AI in roles that can help people. They think of it as we have AI agents that helps the end users. Think of that as your concierge service that's available and that helps you right away. A co-pilot, that is a personal assistant to the human agents, whether it's a support agent or an IT agent or an HR agent. It's like your concierge that's sitting right with you, helping you to solve the problems you're getting right away. And then AI insights, that is helping the leaders get insights into what's happening. So number one is the AI agent automatically handles and repetitive defined tasks, both on the customer experience side and on the employee experience side. It resolves common issues, call it like auto triaging IT tickets or answering simple HR questions or support questions. Think high volume questions that are immediately provide service to the end users. So the driver is we see, for example, we see deflection rates from our customers. They see deflection rates go up all the way from 20% to 60, sometimes even 85%. So the benefit is not necessarily that these questions are deflected, but the end user. Think about the last time you were on a travel website and you were trying to get the information on the phone, whichever modality that you use, whether it's email or chat or phone, sometimes you are on hold and takes time for connect with the agents on the other end. What AI agents is doing, it's giving you that human-like experience. It's improving the customer satisfaction. So that's the outcome that our customers are getting. So we have seen customer satisfaction improve. We have seen deflection rates improve all the way up to 70, 80, 85%. On the co-pilot side, on the other hand, it's augmenting the humans. It's helping the agents respond faster to emails, faster to chats, summarizing the conversations. Otherwise, how has the other agents or other humans have responded to this ticket and it brings that intelligence in. And it helps the IT and the HR agents to complex workflows. And finally, the insights that I mentioned. For the AI insights, it acts as an analyst. It surfaces up the insights from all the signals that are hidden in the tickets, conversations, in the logs, et cetera. Think of it like finding a needle in a haystack. Think of it like a chat GPT for your enterprise. These days, whenever I have a question, I go and ask chat GPT, hey, how do I do this? My daughter the other day asked me about, what is mitochondria? Okay, I'll ask chat GPT, what is mitochondria? It gave me and explained it to me like a five-year-old. That's what our AI insights analyst product does. You can ask any question and it'll give you information about the data that's in your enterprise, whether it's on the IT side or the support side. And something that's unique about Freshworks is that we serve both the customer-facing teams and the internal employee-facing teams. So what that gives us is our AI is sort of a bridge between the external communications and the employee communications. It can be adapted for the IT, HR, and the operations use cases. So it lets unify the experiences and reuse the intelligence. So we call it the connected intelligence. It's a connected tissue across the enterprise, which delivers consistent outcomes across the employee and customer experiences. The outcome is faster resolutions, better experiences, and way more productive teams. And one of the features that you're talking about here too is the AI agent studio that you built. And how did you come about developing that? How did you like test that? It was giving the right answers. Tell us a little bit more about the studio, but what was the process like to actually build that into your platform? Yeah, very good question. I'll start by giving a really good example. I talked to a customer a couple of days back. It's a payments company. And the payments company has grown by over the last two, three years, and they've been using our AI agents. And this is the best compliment that I got. They said, hey, we've been using your AI agents for the last year or so. And somebody tried to hack into your system and they started asking questions about our IP address, how it functions, and it wouldn't answer any of those. It stuck to their guardrails. And they said, I don't know how you guys built it, but your AI agent is way better than the end user who tried to trick them, or the hacker tried to trick them, but it wouldn't trick. Even a human, that's the highest level of compliment that I've seen. So we always start, when we start building the AI agents, we start with a simple but a very powerful question of what does the customer want to achieve? And what we heard was that we want the AI agents to actually understand my business, work across my systems, and save to deploy without needing a huge technical lift. So they don't need a PhD. So that's, so our design principles when we started building this where it has to be no code, low code first, which means the business and operations team should be able to design, configure, and maintain the agents with minimal engineering support. The example that I gave you earlier, this person was running the support center. No technical prowess whatsoever. And this person was able to design and deploy the AI agents. Oh, second is the unified knowledge and context, right? The agents need to access the right knowledge, the data, and the history across various systems, like CRMs, ITSN, HR systems. And these systems, they shouldn't live in silos. So it brings together all the systems. We have integrations with 50 plus external systems. And the third one is multi-channel and multi-modal. What I mean by that is the same agent, once you build an agent, it should work across different channels, such as chat, email, voice. And it has to be consistent and it has to flex between using, that's multi-modal. And it has to. that's multi-modal. And it has to have enterprise grade guardrails, which is governance, security, data residency, and what we call as responsible AI controls had to be built from day one. And when you do all of that is when you get these kudos. So initially we tried this small set of design partners, we cleared up very quickly. Our internal teams are a customer as well. So we have our own IT team, our own support team that uses, and we instrumented everything in such a way that we could see what worked, what didn't, and which helped harden the product. Today, it's AI agent studio is a platform where you can orchestrate agentic workflows across employee experience and customer experience, along with the outcomes and monitoring the performance of how those agents are doing. And what's unique about your agentic workflows that you incorporated into here? So I already touched on three things, right? It's easy to build, easy to use, easy to monitor. That's probably three things I would say, but they're outcome driven. If you look at the other AI agents that are out there, they are like step driven. In our case, you define the outcome, you define the business outcome. The business outcome could be, hey, can you resolve this ticket or provision access to these IT tools, approve a refund. And the workflow orchestrates the steps, the tools, and the human involvement included to get there. And they're cross system by design, which means our agents and workflows can call into multiple systems. Fresh work product, it could be fresh product, which is fresh task and fresh service, or it could be third party tools, so that the automation actually reflects how work happens in real companies, because we've seen that in our customers. They use multiple systems of record, so it orchestrates across multiple systems of record. They're human in the loop, frankly, which means our AI agents support approvals, handoffs, and they can go into copilot mode so that you can blend automation with human judgment in a controlled way. Then governed and observable, which means you can see which workflows are running, what they're doing, how they're performing with the ability to set policies and limits. In short, we're not just automating single interactions, we're automating end-to-end journeys for our customers and their end users, which could be employees or their customers. Early access to our workflows, which we launched in June last year, they report an average deflection of 65%, sometimes achieving up to 85% of the service issues have been resolved by AI agents. PMs make great investors. If you're a product leader curious about angel investing, check out AngelSquad. It's where over 2,000 operators from Google, Meta, and Apple learn to invest in high growth startups alongside HustleFund. I've been a member for years and highly recommend it. They've given me a few 30-day guest passes to share, so head over to go.angelsquad.co slash Melissa, and make sure to act fast as the passes are limited. With these types of AI agents that you're building, it sounds like they could take in an immense amount of data. How do you get started building something that could integrate with all these different tools, have to go through massive amounts of data, all of these different stuff? How did you think about where do we start and how do we test this to then scale it? What was that process like? This is a question that our customers ask is, hey, how do I start my AI journey? It starts with a simple thing. Identify a very simple use case that you want to use for automation, and you can go into the agent builder that we have, and you can describe what you are trying to do. You can give it a persona, and you can deploy. You can handle simple use cases to begin with. As the simple use cases, as you get more confident that simple use cases are getting answered, you can slowly scale it to other use cases. Because as I said, it's tightly integrated with the human loop. For simple questions, the AI agent can answer, and you can instruct it to answer simple questions. For more complex ones, you can go to a human agent, so the human agent can answer the questions. There is this virtuous loop of human. We call it people-first AI, which is why the people-first AI, the people make decisions and people are driving AI to be more effective according to how their pace of AI adoption should be. I love that. I love that. People-first part of AI, because everybody always thinks it's just going to replace them all. But yet, it sounds like you have a lot of people who are actually using this. You're augmenting people to be able to deliver this support with AI. That's absolutely right. Yes. We have hundreds of customers that are using it. Every one is unique. We have customers such as Gayle's Bakery, which is an SMB customer to a very large customer that I mentioned at Payments Giant, iPostal. We have multiple different customers of different sizes that are using it. Everybody is in their own journey. Some are adopting it really fast. Some are cautious about it. Some want to show me before I use it more. You have companies all the way from Gayle's Bakery to Seagate to Databricks, all the customers that you mentioned, larger companies that are using it and smaller customers that are also using it. What we believe in is, like I said, it's people-first AI. People decide which use cases you use it for, what pace at which I want the service delivery to be automated, and which places I want to be automated. One last thing, I'll end by saying our motto is we deliver complicated solutions to businesses of all sizes. Yeah. Interesting because it's such a span, right? You're a small business or what you've been mentioning here versus your Databricks, versus these large corporations. Where do you think the uncomplicated part comes from? What do you really focus on there to make sure that it can span everything but still be powerful? It starts with the user. Remember, if you are ultimately a user, whether you are in Seagate or Databricks or whether you are in Gayle's Bakery, the end user wants an uncomplicated system all the way, like I said, three things, to setting it up, to using it, and then monitoring usage. Think about it this way, right? Formula One or any other car, what we're doing is we are giving the car to the customer to use and it's super simple to use. They can use it however they want, whether it's a large customer or a small customer. Whichever size of the customer actually is, it doesn't really matter. And this is where we see some of these larger companies that have this bloat. We need to have consultants come in and set it up and it takes a ton of time to get up and running and it takes weeks of training. So none of that exists for us. You can just get it up and running in a few hours. I love that. I love the simplicity. And I think it's so nice, especially for larger companies too. I always get into, I feel like, arguments or questions from people that are like, hey, if it's just an internal tool, does the user experience matter, right? Like if my employees have to use it, my customers aren't seeing it, do I really care how easy it is to use or how nice it looks for them? And you're making an experience for those employees and you're focusing on the simplicity, which I think is amazing. Absolutely. If you ask any IT team in the world, who's your customer? They would say, my employees are my customers. And we have our own internal, our own CIO that says, my CSAT is driven by your satisfaction. I'm an employee and I'm his customer. So he says, as long as my customers are happy, that's what drives their customer satisfaction. And same thing applies to support centers on the CX side. Customer satisfaction is the most important thing, whether it's internal or external customer. And Srinivasa, when you were encouraging your product managers to maintain that customer centricity, especially at the rate of everything involving, like we've got a million different AI tools coming out, all this new technology, the market's moving really fast. What types of things are they doing to make sure that they don't lose sight of that customer? I think the number one thing is they need to be relentlessly outcome focused. I encourage all my product managers to talk to customers. I personally talk to at least five to 10 customers per week. And that's like at least a third to half of my time every week is talking to customers. Ultimately, the product manager's job is to be the champion of the customer to the internal teams, whether it's engineering, support, professional service, any of those. So stay close to the customers, understand their world and connect everything that we do, whether it's strategy, roadmap, design, or AI, and backed by the real outcomes that we deliver for them. And the tools and trends will change. Like today we have Sigma make tomorrow could be something else, be lovable, or cloud code, all these things have means to an end. But the mindset of being outcomes driven and being customer obsessed will always be relevant. I totally agree with that. Srinivasa, my last question for you, if you could go back early in your career and give yourself one piece of advice, what would it be? That's a great question. Because I ponder about this all the time, but I have young kids and I want to give them advice. So the first thing I would say is fall in love with the problem, not your solution. And learn the language of business early, revenue, margins, unit economics, so that you can connect your product decisions to business impact. The third is invest in storytelling and influencing, not just analysis. Great ideas need great champions. And finally, be comfortable with ambiguity and change. The best opportunities often look messy at first, and maybe the most important, enjoy the journey. Product management is a team sport. It should build matter as much as the product situation. I think that is fantastic advice for people out there. And I really love your connection of the business language for product managers. It's been a hot topic lately. Thank you so much, Srini, for joining us. If people want to reach out and follow you, what's the best place to go? They can reach out to me on LinkedIn. My handle is Srinivasan28, that's S-R-I-N-I-V-A-S-A-N-28. And my Twitter handle is S-R-I-N-I-Y-E-R on Twitter. Those are the two places they can reach out to me. Okay, great. And we will put those links on our show notes at productthinkingpodcast.com. Thank you so much for listening to the Product Thinking Podcast. We'll be back next week with another amazing episode. Make sure that you like and subscribe so that you never miss out. And we'll see you next time.