Keeping it uncomplicated: A conversation with Srinivasan Raghavan
Freshworks’ product chief shares his vision for speed, simplicity, and responsible innovation
Just a few weeks into his role as chief product officer at Freshworks, Srinivasan Raghavan was put on the spot by CEO Dennis Woodside during a quarterly all-hands meeting: How was he planning to guide product development at Freshworks?
Raghavan leaned on a familiar analogy: race cars. “In an F1 race, when a car pulls into a pit stop, you have 10 people touching the car at once, each with a critical role,” said Raghavan. “That’s how I see my job: making sure our product moves faster, with every team working in sync to keep us ahead."
In a Q&A that included questions from Woodside and employees, Raghavan discussed his product development strategies, implementing responsible AI, and lowering costs. Here are some highlights from the conversation.
How do customers describe the pain points they’re experiencing with incumbent solutions?
The biggest complaint about incumbents is that they're bulky and complicated. Just getting value from the product takes 12-18 months. When you look at big brands like ServiceNow or Zendesk, the implementation alone can take 12-18 months.
Read also: The new architects of employee experience
Our customers appreciate the simplicity we bring. But I'll self-reflect, too. I get called into two kinds of meetings. About 20% are prospect meetings where I sell our vision and roadmap. The other 80% are escalations where someone forwards an email about an unhappy customer.
That's the part of the CPO job I love: It's not glamorous, because you get yelled at by customers, but you learn a lot. I'm discovering the gap between customer expectations and what we deliver. Sometimes things don't work as advertised, which happens, but we hold ourselves accountable.
The second issue is that customers expect more quality. They're paying for a BMW and expecting a BMW product. Many features get built because customers request them. The key learning is to bring this feedback to our product and engineering teams to improve what we offer, which ultimately helps us win the next big customer.
How do you keep product development flexible against a changing AI backdrop?
We've built an architecture that allows us to swap in and swap out multiple models. Even before recent developments, we used a combination of Microsoft’s OpenAI and Anthropic’s Claude. We've also been exploring other models, like Mistral.
What's fascinating is how the cost curve is evolving. I've seen this before with automatic speech recognition (ASR). Five years ago, ASR cost $1 per minute. Now it's dropped by 10,000%. We're seeing the same pattern with LLMs: Costs are coming down rapidly while performance increases.
Our thesis is that the world is evolving toward AI agents and co-pilots. Insights will probably become part of powering these co-pilots and AI agents. What we're really doing is going after a completely new market. Customers spend approximately $10 on labor for every dollar of software, so there's huge potential for efficiency gains.
How are you ensuring AI implementation is safe and trustworthy for customers?
Customer concerns typically fall into three buckets:
Our customers appreciate the simplicity we bring.
First is data residency, especially for customers in regions like Europe who have specific privacy and compliance requirements. Second is data security: Customers want assurance their data isn't being used to train models. Third is responsible AI: preventing data hallucinations and inappropriate responses.
We're addressing these through multiple approaches. For data residency, we have guardrails in place. For security, we work with our vendors, including Microsoft Azure OpenAI. For responsible AI, we're investing in solutions to reduce hallucinations and set governance policies around LLM responses.
How do you ensure your product remains uncomplicated, especially as AI-driven capabilities expand?
Most companies, whether SMB or enterprise, use about 20% of available features. Our promise is to give you a simple product that meets your needs at a price point you're willing to pay. This applies across the board—SMB, mid-market, or enterprise. Everyone wants to save money and get value from their investment.
We know the interface and user experience will change significantly with AI. Today, 50% of customer interactions involve human intervention. In 12-18 months, that could change dramatically. The key is ensuring our AI remains humane and the interface is customized for each industry vertical and user type. We need to deliver personalized experiences while staying true to our promise of simplicity and value.