Beyond engineering, every department is using AI to drive efficiency at Freshworks: CEO Dennis Woodside

Woodside shared with Moneycontrol Freshworks’ AI journey, financial discipline, and what comes next.  Edited excerpts of the interview:

You have increased your revenue guidance for the year, what's driving this confidence?

Yes, we continue to see strong growth. The IT service management side of our EX business, which primarily supports IT teams, is growing fast. That segment is gaining traction with larger accounts, giving us confidence going forward. This past quarter, our EX business grew 33 percent year-over-year in constant currency. The CX business grew 7 percent, consistent with the prior quarter. We're increasingly seeing larger accounts moving over to us from incumbents like ServiceNow and Salesforce – companies that offer heavyweight, complex platforms. Customers want something easier to use but still enterprise-grade, scalable, and powered by AI. AI has helped here – a large portion of our new and upsell deals now include an AI component.

There’s been an improvement in your operational efficiency too. What’s driving that internally?

We’ve improved our cash flow margin substantially – moving from single digits to guiding over $200 million in free cash flow for the year. We hit 30 percent operating margin this past quarter. That’s strong bottom-line performance. Internally, we’re being smarter with expenses – investing where it counts and cutting inefficiencies.

AI plays a huge role here. Across the company, we’ve implemented over 70 AI applications. For example, our engineering teams use GitHub Copilot and Cursor for coding, and Figma for AI-assisted mockups. Our support teams rely on Freddy Copilot and Freddy Self-Serve. Our billing and legal teams also use Freddy to answer queries more accurately. It’s all about scaling efficiently while improving outcomes.

Do you have a number on how much of your workforce is using AI in day-to-day work?

Everyone writing code is using AI in some form. Cursor is great for generating first drafts from English prompts, while Copilot is better for modifying existing code.

Even beyond engineering, every department is using AI. Our recruiting team built its own AI tool – with engineering support – to sift through 10,000+ job applications we get each week.

We’ve also built internal tools to analyse our cloud costs, which has helped optimise spending. Some of this internal innovation is evolving into market-facing products – like Freddy AI Insights, which gives conversational access to business performance data.

Internally, how are you using AI to drive efficiency? Any metrics or dollar investment figures?

We don’t track AI investment by dollar amount. Instead each function must identify at least one AI application they’ll use regularly to improve work. We encourage experimentation. For example, engineers use GitHub Copilot, designers use Figma’s AI tools — we see this driving significant long-term value.

Let’s talk demand pockets. You mentioned a shift from SMBs to larger customers. Where’s the focus today?

SMBs still matter – they make up 39 percent of our ARR. That segment performed strongly this quarter thanks to improved marketing and sales alignment. But yes, we’re also going upmarket. We recently won deals with Travis Perkins in the UK, a large US hard drive manufacturer, an optical networking company, and even a division of the US government. All of them had been with ServiceNow for 10+ years. Our product has matured significantly in the last few years and we’re now better equipped to serve large enterprises.

Your GAAP net profit almost turned positive this quarter. What’s the outlook there?

Yes, we reported a GAAP net loss of just $1 million this quarter – a big improvement from previous quarters. That said, profitability may fluctuate due to seasonal expenses like bonuses. But we're getting close to consistent GAAP profitability. On a cash basis, we’ve been profitable for a while. We're guiding over $200 million in free cash flow for the year. That gives us flexibility to invest in products, people, and shareholders.

You’re sitting on a healthy cash balance. What are your top priorities for that capital?

There are four key areas. Our people — by creating opportunities and career paths internally. Next is product and customers. All our R&D is in India – Bangalore, Hyderabad, and Chennai – and that’s where we continue to invest. And, acquisitions. We’ve done small acquihires and bigger ones like Device42 last year, we will continue to look for opportunities here. We will also do buybacks. We repurchased over $100 million in stock in Q1 as part of our $400 million authorisation.

Let’s talk AI investments. Are you spending more on talent or infrastructure (LLMs, GPUs, etc)?

Primarily on people. We have over 1,500 technical staff in India, many with 20–25 years of experience in their fields, focused on AI products. We’re also working with nearly 40 different LLMs – some built in-house, others open source or via providers. Despite this investment, our business remains profitable. In Q1, we had 2,700 Copilot customers (up 500), 1,500 AI Agent customers, and 500 in beta for AI Insights. The value is clear – Copilot users see 30 percent productivity gains, while AI Agent users see 50–70 percent query deflection. That said, we’re still early – we have 73,000 customers in total, so there’s a long runway ahead.

Your Net Revenue Retention (NRR) has been around 104–105 pecent for the last few quarters. However, it has not improved significantly. Do you still see churn in customers?

Churn is improving steadily. But at the same time, companies are hiring fewer people, so seat expansion is slower than before. We’re still adding seats, just not at the previous rate. That’s why NRR has remained flat.

Are you experimenting with AI pricing? How are you thinking about monetisation?

Yes, AI pricing is still evolving. We’re trying various models. Freddy Copilot is a seat-based add-on at $29/seat. Freddy AI Agent is usage-based – you buy sessions. Device42 is priced per device/laptop. We’ll keep testing what works – what to include in base plans vs. as paid add-ons – and adapt based on market feedback and customer value.

Are you able to monetise AI now or is it still an add-on?

We’re monetising Freddy AI (our frontline AI agent) on a per-session basis. For example, PhonePe pays us for each session its users interact with the AI, which helps them avoid costlier human interactions. Copilot is monetised at $29 a month on top of the base licence. In Q1, 1,500 customers paid for Freddy AI and 2,700 for Copilot. It's now a core part of our revenue, with about half of all large new deals including Copilot.

How is Freshworks navigating global macroeconomic uncertainty, especially in the US?

We’re a global company with 45 percent ARR in North America, 40 percent in Europe and 15 percent elsewhere. We're not tied to any one industry or government contracts.

In Q1, we saw no major change in buying behaviour from Q4—churn, expansion, and pipeline remained stable. Our products automate IT and customer support, which are essential, not optional. Plus, we offer modern, AI-enabled solutions at half the cost of legacy platforms like ServiceNow or Salesforce.

You started restructuring after joining. Is that phase over or will more optimisation happen?

We're now in a hiring mode, especially in AI and engineering roles across Chennai, Hyderabad, and Bangalore, and in go-to-market functions. We’re investing in the future. If the market slows, we may slow hiring but right now, it’s a growth phase.

It’s been a year since you took over as the CEO. How has the experience been and what are your priorities now?

It’s been amazing, especially spending time in India — I've visited four or five times, met the team, and engaged with customers. We've moved upmarket significantly; 75% of our EX business now comes from mid-market and enterprise clients. AI has gone from whiteboard to revenue in under a year and will be central to our sector's future.