In this episode of AI at Work, I sit down with Dennis Woodside, CEO of Freshworks, to uncover how real companies are getting true value from AI.
Dennis shares how Freshworks has built AI tools that help businesses resolve routine questions automatically, boost agent productivity, and give managers clear performance insights without needing complex dashboards. He explains the company’s focus on making AI quick to deploy and simple to buy, so mid-sized companies can see immediate returns without endless consulting bills.
We explore customer stories like Total Expert, which saved thousands of agent hours and saw a 250 percent return on its AI investment. Dennis also talks about the lessons learned from integrating AI internally and how the company stays flexible enough to adopt the latest advances from across the industry.
This conversation is for anyone who wants to see beyond the AI hype and hear how smart companies are using it to save time, cut costs, and let people focus on more rewarding work.
[00:00:03] Welcome to AI at Work, a podcast which is part of the Tech Talks Network. And in this podcast, we're going to venture into the transformative influence of artificial intelligence inside the workplace. And our discussions will focus on both the remarkable breakthroughs, but also the complex challenges of integrating AI into our everyday business functions and
[00:00:30] workflows. Well, today I'm joined by Dennis Woodside. He's the CEO of Freshworks. And together we're going to unpack the real world impact of AI on everything from efficiency, customer experience, and bottom line results. Because Freshworks has been at the forefront of democratizing AI for mid-market businesses and ensuring that cutting edge enterprise grade AI
[00:00:55] tools are actually accessible without the hefty price tag. So I'm looking forward to discussing how companies can track and prove ROI in AI deployments, and also explore why mid-market businesses shouldn't be left behind and understand what exactly leaders need to be focusing on when integrating AI AI into their daily operations. So if you're wondering how to implement AI at the speed of innovation,
[00:01:24] while also ensuring it actually delivers tangible business value, this episode is packed with insights you're not going to want to miss. But enough from me. Let's get Dennis onto the podcast now. So thank you for joining me on the podcast today, Dennis. Can you tell everyone listening a little about who you are and what you do?
[00:01:45] Great. Well, first of all, thanks for having me, Neil. I am the CEO of Freshworks. We are a leader in providing AI-powered software that makes IT and customer support teams more efficient. I'm a husband, a dad of two almost grown kids. I'm a triathlete. I've done the Ironman triathlon 17 times. I grew up in Philadelphia, Pennsylvania. I live out in California now.
[00:02:10] I've spent over 20 years in the Valley working mostly with disruptive tech companies. I joined Google when it was a small company at about a thousand people and had roles in London for about four years with Google, ran sales in North and South America, ran a company that we bought. I was the chief operating officer at Dropbox and helped take that company public. So I like working for kind of the disruptive innovators and really Freshworks has that DNA, which is why I joined a little over two years ago.
[00:02:39] Well, thank you for taking the time to sit down with me today. And one of the reasons I was excited to get you on the podcast, there seems to be a real feeling around AI at the moment. We've all seen the hype over the last two years. But as that matures, many businesses are struggling to find the ROI in their AI project. They're struggling to find that measurable impact that it's having on their business. But then I arrive at Freshworks and you've showcased efficiency gains through AI.
[00:03:05] So can you share an example or maybe a case study that highlights some of these gains and how they've translated into a real financial impact? Because I think this is a story we don't hear about enough and something a lot of business leaders are looking for guidance with right now.
[00:03:20] Yeah. So I would start by just saying that in the software space, one of the things that we see is complexity really overwhelms the business owner in trying to drive innovation and change. Complexity leads to inefficiency, higher costs. It slows down decision making.
[00:03:42] We hear this all the time that our competitors provide solutions that are highly complicated, in some cases bolted together multiple different products. And that leads to them not getting the value that they thought they were buying when they when they made the decision to go with the vendor. So they have to hire a bunch of consultants to modify the product to suit their needs, which just makes the problem worse.
[00:04:04] You then introduce AI to that and it becomes very hard to get value from from a technology that we all know could have a lot of promise. We all use ChatGPT in our personal lives now. We all see how that has helped us become more productive. And the question that we get asked by customers is why can't it be that easy? Now, our business was built on the premise that that you can get value from our products very fast.
[00:04:29] We launched our first product in 2012, and that was really focused on a small business. It doesn't have a lot of resources and needs to get a customer support help desk up and running really quickly. We've taken that same ethos as we've grown over the last decade plus to larger and larger companies.
[00:04:50] So today we have 72,000 businesses that are using us for either customer support or IT primarily big companies like DHL, Amex Business Travel, Bridgestone Tire, Sony, Nucor Steel, Airbus all rely on us. And increasingly, they're looking to bring AI solutions in to make their agents more productive and then to deflect rope queries so that the agents don't have to handle those things.
[00:05:16] And we're seeing across the 70,000 customers, those that are using AI are seeing about a 30% improvement in the productivity of their reps. And those who are using our frontline L1 AI agent, that's the product that deflects tickets, they're seeing between 50% and 70% deflection rate. So they're getting real value because we built the product so it's fast, fast time to value, easy to use, easy to deploy.
[00:05:41] And as we're saying here, business leaders all around the world are facing this intense pressure to justify their AI spending with measurable returns. So just to dig a little bit deep on some of that that you just said there, can you expand on Freshworks AI products and how they're creating both for customers and indeed internally too? So our AI strategy is premised on three distinct personas or use cases.
[00:06:09] The first is that L1 agent, L1 support. So think the product that we have is called Freddie AI agent. You can deploy it in email or in chat and it answers questions based on your product manuals, your existing FAQs, your history of interactions with customers. And a lot of those kinds of questions that we get, that our customers get, are rope queries that they really don't need an agent necessarily to answer. But it's important to answer them accurately and quickly. And that's what the AI does.
[00:06:38] We have around 1,300 customers paying us for that AI agent product today, ranging from small customers all the way up to large customers that are deflecting literally millions of queries every month. The second persona or use case is the CoPilot product, which we call Freddie AI CoPilot. That's to improve the productivity of an agent themselves. So think about a customer bar like S&P Global.
[00:07:04] S&P Global serves brokerages and financial institutions around the world with a complex set of data solutions similar to a Bloomberg terminal. The questions that the S&P Global gets from its customers are very complex technical questions. And it's very hard for an agent to know the answer to those questions off the top of their head. So historically, the first thing an agent would do would be research in the product manuals.
[00:07:33] What product are they using and how do I solve this problem for the customer? That can take an hour. With AI, AI already is trained on all that material and the AI suggests an answer to the agent instantly. So you have an answer that's likely to be much more accurate than what the agent would deliver much faster. It's good for the customer or our customer because they get more productivity out of that agent. And it's good for their end user because they get a better answer faster. So that's co-pilot.
[00:08:00] And then the third persona and product is called AI Insights. And that's designed for the manager of the team. So think of a manager of an IT team who historically would come in on Monday morning and look at a bunch of Power BI dashboards or something similar to try to understand what's going on and try to infer where might there be problems, where my attention is required, where are we seeing good performance as well? You don't need, in the age of AI, those dashboards at all.
[00:08:30] You can just have a conversational interface and ask questions of the AI. Share with me any interesting performance trends that happened over the last week. Spikes in inbound questions, changes in CSAT trends. All this stuff is answerable because the AI has been trained on your service environment and is constantly up to date. So those are the three areas that we're driving. We're seeing traction in the first two are in GA.
[00:08:58] We have combined over 3,000 customers paying for those two products. We just launched AI Insights into beta and we're excited about bringing that into GA later this year. If we take a look at any of those organizations, many routine tasks will be frustrating for both employees and customers alike. And ultimately, they also end up holding back innovation. And I suspect we've both seen that firsthand too. So I'm curious, which of these kind of tasks have you seen AI transform?
[00:09:28] And what impact has it had on the overall productivity? You mentioned a great stat a moment ago, but I'd love to dig a little bit deeper on that. Yeah. So let me give you a specific example. We have a customer of ours that is called Total Expert. They're a customer agent or engagement platform that's for the financial services industry. So their customers are financial service providers and things like CFPs.
[00:09:56] So they had lots of questions about their product coming from those analysts and those end users of their products. And they implemented our AI co-pilot as well as AI agent because they didn't want the kind of lower value questions, the rote questions that they get all the time to go to a human. It's not a good use of the human's time. They want the humans focused on much harder problems where empathy is required, things that they haven't seen before.
[00:10:25] So by implementing AI agent to deflect those questions and then also implementing co-pilot to assist the human agents to be more productive, they saved about a thousand hours of those agents. They saw over 250% ROI on the investment, on the cost that it took to deploy those AI agents, something like over $100,000 in annual savings. And they're seeing about 23% of their messages that are now resolved by AI agents.
[00:10:55] And they just got started. As the AI gets better, as they get better, those numbers are just going to go up. So that gives you an idea of what – that's a huge transformation in that company's interaction with their customers. It frees up a huge amount of time for those agents to do more creative things, more interesting things for the company. And it really does unlock innovation within that organization because people who were deployed on things that were not innovative, not creative, just wrote questions.
[00:11:23] That's no longer the case. They can be used – their minds can be used in a much more creative way to solve bigger problems. And, of course, implementing any new technology, especially at the speed that it is advancing right now, can be challenging too. So what kind of strategies do you employ at Freshworks to deliver both those immediate and low-hanging fruit gains, but also long-term returns on those AI investments?
[00:11:51] Because we go after those short-term gains to begin with, but it's those long-term returns that are equally as important, right? Yeah. So we – a couple of things that are unique about Freshworks. We very much are focused on what we define as the mid-market set of customers. So for us, that's a company with 20,000 or fewer employees. So why is that the segment that's important for us? Well, that's where you have sophisticated IT or customer support needs.
[00:12:20] You have hundreds of agents. Typically, you have global operations. You have regulatory constraints. You need a sophisticated platform, a sophisticated solution for IT and for customer support. It needs to be AI-enabled across the board. At the same time, those companies don't have the resources to hire 10 consultants for a single product. You hear this all the time. I think I saw a stat that PwC has a multi-hundred-million-dollar business just in serving ServiceNow customers.
[00:12:49] We hear this all the time. You don't – the mid-market does not have the resources to go hire 10 consultants to tend to babysit a piece of software that they bought. They want the software to work out of the box. That's what we've built. So from the beginning, the software is designed to get up and running very quickly. We have a large retailer we just signed up last month in the U.S. with thousands of stores around the U.S. And they were up and running on our customer support product moving off of Zendesk, one of our bigger competitors.
[00:13:19] They were up and running in two weeks. And in enterprise software, that's unheard of. Usually, you're talking months to deploy a piece of software. You can be talking years depending on what solution you're dealing with. We built the product so it's fast time to value from day one. We built the AI in the same way. If you want to use our AI agents, you go to a screen and you point it at the data that you wanted to learn.
[00:13:46] It could be web pages or PDFs, product manuals. It understands that information in a matter of minutes. And you can then deploy that AI agent on any chat surface that you might have in your email inbox so that customers who are asking questions about that content, they get answers very conversational way directly from the AI. So we have customers that are literally up and running in an hour with that product.
[00:14:16] And again, that's unheard of. If you look at what's going on, what Salesforce has announced, what ServiceNow does, they require a lot more lift to get value from their AI solutions. Co-pilot, you turn it on and it's on. It's available for all your agents. Your agent is in the same screen that they're used to working in every day. They see an icon that indicates that Freddy AI is enabled. And when they tap that icon, it suggests answers. It summarizes answers. It enhances tone of emails going out the door.
[00:14:46] It just works out of the box. So it's actually very hard to build a product that's that uncomplicated. You have to really think about the problem from the customer's point of view, from the admin's point of view. And that's what we've done from day one because when you're focused on a smaller business to start with, even though we have large customers now, the small business, you need to show value immediately.
[00:15:10] And before you came on the podcast today, I was doing a little research and I was reading that you're incredibly passionate about democratizing AI for those mid-market companies and making it more accessible. Because I know a lot of them struggle. So I'm curious, from everything that you're seeing, from the trends you're observing, from the conversations that you're having, what are the key barriers that these companies are facing in integrating AI? What's holding them back? And how are you addressing some of these challenges?
[00:15:36] Well, the first is I talk to CIOs and execs and IT all the time. Most mid-market companies are already seeing a lot of experimentation in AI within their organization. Like at Freshworks, for example, we have 15 different AI projects that are going on across engineering, product development, design, marketing, sales, customer support, you name it. Some involving our products, some involving third-party products.
[00:16:06] So every decision maker knows that AI can help their business. They're just trying to figure out how. So what are they looking for? Well, they're looking for something that they can get up and running fast because they want to be able to A-B test quickly and see what they're getting positive results from and what's working. So they have to be able to turn it on quickly. To my point earlier, without a whole lot of hassle, without paying somebody to do an integration that should come with the product out of the box.
[00:16:34] They want it to work for their agents. And they want their agents to see that this is actually a way to improve outcomes, improve quality, and enable the agent to work on things that are more interesting and have greater value to the company over time. And they don't want surprises in the process. They don't want to be charged for things that they didn't know they were going to be charged for. They don't want to wind up having a bunch of issues with how their data is treated.
[00:17:02] They need to ensure and know that their data is going to be handled securely. All of those things are what we talk about with customers when we're getting a company up and running or talking about introducing a new company to Freshworks. And those are the things that we're seeing as super important for those decision makers. And I think there is a fear around many medium-sized businesses that will eventually get priced out of the AI game or a perception that they could.
[00:17:26] So how can enterprise-grade AI tools be made more accessible, more affordable for this mid-market organizations? And what role are you trying to play at Freshworks in helping to level that playing field too? Well, so our products are – the pricing is on our website. It's completely transparent. And when a customer comes in and wants to try Copilot, for example, they don't need to provision their entire organization.
[00:17:53] They can provision 10 agents, see how those 10 agents do, compare their productivity, customers' CSAT, the accuracy of answers, compare the 10 that are using a Copilot with 10 that aren't. And then it's a rational economic decision. Is it worth paying the cost to get access full-time to the AI for the whole team? So that's one way that we make it easy to evaluate, easy to use. They don't have to sign a year-long contract to evaluate the product.
[00:18:22] Our AI agent, which is, again, the L1 support agent that deflects queries to begin with, we charge on a consumption basis there. So every interaction there's a charge for, but the customer support team or the IT leader can very clearly compare that cost, which is actually quite low. It's around $0.10 in the interaction. Compare that cost with what it costs them on average to answer a question.
[00:18:47] Most support teams will say, on average, it costs you easily $10 in terms of the support agent's time, amortized over the time it takes to answer the question. Easily $10 a resolution or charging a fraction of that. So we believe in being really fair, transparent around pricing. That's an area where a lot of our customers get hung up on with respect to our competitors.
[00:19:12] I was also reading before you came on the podcast, in your first year as CEO, if you reflect on this first year as CEO, what have been your most significant lessons learned in integrating AI into Freshworks operations and that overall strategy? Because the speed of technological change is just ramping up at the moment. And so much has happened in a year and within the last couple of years as well. So what are the biggest lessons that you've learned, would you say?
[00:19:41] Well, I would say, first of all, Girish Mahathabrutin, who's the founder of Freshworks, and he preceded me as CEO. He has been thinking about AI for a lot longer than most. We first had an AI-enabled product launched in 2018 at Freshworks. When ChatGPT first kind of came out, he jumped on it as quickly as we could. We had our first co-pilot product in beta about a year and a half ago in GA, a little over a year ago.
[00:20:10] And we went from no customers paying for AI to, like I said, 3,000 plus customers paying for AI and getting value from AI today from us. So I think that we are able, as a company of our size, we have the heft and size to deliver amazing products. But we're not a gargantuan tanker, oil tanker that takes a long time to turn. So we were able to move really quickly and deliver on AI.
[00:20:37] And I thought that was really important for us to do because our products are as competitive as what you're seeing from a service now or a Salesforce companies with 20 times our market cap. And are providing the kind of resolution, the kind of impact to our customers that our customers expect, which is amazing. I think internally we see some of the same things that our customers do, which is, you know, we want to encourage the use of AI in every department.
[00:21:03] But we also need to control for quality, for security and for cost. And so we have taken an approach to be very clear about, OK, what's an experiment and what are the gates for taking an experiment to scale? A, B, testing any kind of new product that comes in. We encourage teams to go out and look to see what's out there, not just in IT.
[00:21:25] But if you're on the sales team, we want you to be playing around with other tools that are out there that might enhance your performance as a seller. And when those tools get to a certain a certain point, we'll do a test and see if they do, in fact, improve the output of the team. We have a couple of tools that are used as an example in prospecting that have been quite, quite productive in composing emails in a way that's that's highly relevant for the customer we're reaching out to.
[00:21:53] We've we've used a product that helps an account executive prepare for a meeting with a company by summarizing all the relevant data about their IT needs, about their customer support needs and about their business. Very simply in human language that the A.E. can read that could typically take an hour or more of time historically for them to do themselves. So that's the kind of experimentation we want to encourage.
[00:22:19] And then once we prove to ourselves that there's real return from those investments, we'll scale it out to the whole organization. We'll train the organization on how to use the A.I. to enhance their their job performance, to make them more productive, to take some of the the unfun work that everybody has out of their day and speed it up and make sure that they can spend more time with their customers.
[00:22:46] And as we become outside of the hype cycle that surrounds A.I. and we're seeing a lot of maturity now, we're getting back to solving real problems, measurable impact, ROI, all that more interesting stuff. So if we were to look ahead now, we have this maturity. How do you see the role of A.I. evolving in the industry? And are there any other market trends that you're monitoring closely or you just believe will help shape Freshworks strategy over the next coming months and years?
[00:23:17] Well, I think it's a it's a really exciting time. It's it's it's similar to when. Every company was realizing that mobile and mobile computing was going to change their relationship with their customers and they're trying to figure out, well, how do I how do I play in that space? It's also giving rise to completely new business models, much the way mobile gave rise to services that we have now that we didn't have 12, 15 years ago. We couldn't even dream of.
[00:23:42] And I think we're in the early stages of understanding how can A.I. improve our work lives, improve our lives at home. So it's always exciting to be in that space. That's what motivates us. That's what motivates our engineers. For a company like us, we also can take advantage of the entire industry's innovation. We've built our product so that we can connect any large language model that pops up and they're popping up all the time.
[00:24:10] So an example we know for coding today, Anthropic has a superior product to everybody else out there. So we can tap into Anthropic for coding needs and any kind of product that we want to build that that that touches a coding use case.
[00:24:26] But we also know that Google has a very good solution for Gemini for for processing imagery and doing things called multimodal, where a customer or an employee sends a screenshot to the A.I. and says, which is which says something like this is my I got this problem screen. What do I do next? The A.I. can actually interpret that because it understands images.
[00:24:49] So by architecting our products so that we can take advantage of the industry's innovation, we think we're going to constantly be able to stay on the cutting edge. And when we talk about like A.I., like Copilot, it's not one product. It's really 70 different use cases like summarize an article, enhance the tone of an article, improve the actual grammar of the article.
[00:25:10] All of those things are different use cases where as the industry evolves, there are going to be different LLMs that solve for those use cases differently. And that's how we're sort of staying on the front of innovation. And I think as the industry innovates, our job is to bring that innovation to our 70,000 plus customers in a way that's secure, fast time to value, integrates with their existing solution, accessible for the agent, accessible for their end customer. That's that's what we can uniquely do.
[00:25:39] And that's that's how we're building our business there. Well, I'm very conscious. We've been incredibly forward thinking today, looking into the future, how we can prepare for that. And right here in 2025, you are the CEO of Freshworks. You've enjoyed a fantastic, hugely successful career. But of course, none of us are able to achieve any degree of success without a little help along the way. Sometimes there's someone that sees something in us, invests a little time to us, don't always get the recognition they deserve.
[00:26:08] So is there somebody that you're grateful towards that maybe saw something in you many years ago that we can give a little shout out to, a little thank you to today? Yeah, I think so. So I think I call out two people. One is Samir Gandhi. Samir is on our board. He is a principal investor at a firm called Excel, famous, well-known venture capitalist in the valley. He and I worked together first at Dropbox, where he was also an investor. He's invested early in some of the biggest names in technology.
[00:26:35] He found Geerish or Geerish found him over or almost 15 years ago and really funded this dream that Geerish had of building software that's accessible, that's uncomplicated, that works out of the box for its customers. He made that about 15 years ago and here we are. So he's always been super helpful to me, sees around corners type of person, very visionary, pushing forward. And then the other person I would call out is Shona Brown.
[00:27:03] Shona and I have worked together for, at this point, probably 20 years. We were at McKinsey together. We were at Google together. She's now the independent chairperson of Atlassian. So I see her regularly and it's great to kind of get a sense for her as to what she's seeing and how she looks at scale. She's very good at scaling organizations and scaling teams. So those would be two that I would call out. Oh, fantastic. Absolutely love that. A big thank you and a shout out to both of those.
[00:27:30] I think I'm at the Atlassian event in a few weeks. Maybe I can deliver that message purposely, personally. But for anyone listening wanting to dig a little bit deeper on all things Freshworks, anything we talked about today, some of the AI stuff too, where would you like to point everyone listening? Yeah, I would just direct them to Freshworks.com. You can find everything out about our products there. If you're interested in a job there, you can find it there as well.
[00:27:56] All of our AI announcements, you can get case studies on hundreds of our customers as well. And then if you are interested in kind of staying abreast of us generally, if you can find us on LinkedIn, either me or Freshworks, you can follow us and we're constantly publishing interesting content there. Well, as we said at the beginning, there is so much hype around AI at the moment.
[00:28:18] But focusing on real world gains, focusing on metrics that matter, delivering both immediate, long-term returns on AI investments. That's the magic. That's the stuff that excites me. And hopefully everybody listening too. So I would urge everyone listening to check out those places. I will add links to everything you mentioned to make it easier for them. But more than anything, thank you for starting this conversation today. Thank you, Neil. Appreciate it.
[00:28:43] Big thank you to Dennis for giving us a practical playbook for ensuring that AI isn't just a hyped investment, but a business multiplier. One that reduces costs, improves customer service and drives efficiency in measurable ways. And a few of the things I'll be taking away is that AI doesn't have to be complex or exclusive to the biggest players.
[00:29:08] Mid-market businesses can harness AI effectively with the right tools, transparent pricing and a strategy that's focused on impact rather than just experimentation for the sake of it. So thank you, Dennis, for breaking down how AI is leveling the playing field for mid-market companies. But now over to each and every one of you for listening. What's your biggest challenge improving ROI on AI?
[00:29:37] Are you seeing tangible efficiency gains on your investment right now? Or is AI still more of a concept than a reality in your business? Let me know. LinkedIn, X, Instagram, just at Neil C. Hugh. Send over your thoughts and we'll keep this conversation going. And then hopefully we can all get ROI from those big tech investments. But I'm afraid we're out of time now. So I'll return again tomorrow with another discussion. Hopefully you'll join me again then.
[00:30:06] Bye for now.