From 135 Clients to 3: Talvin.ai Pivots Right

He traded 135 clients and a 50-person agency for 3 customers.

Most people would call that failure. I got off the call thinking: "This is exactly the kind of founder I want to back."

I recently caught up with Newan, the founder of Talvin.ai. His story is the perfect antidote to the "raise money on a slide deck" culture we see so often in AI right now.

Here is the breakdown of our call, the "agency trap" he escaped, and the three green flags that make his company one to watch.

The Backstory

Newan started his journey at 19 years old. He literally finished his A-level exams on a Wednesday and started his first company on a Thursday.

I’m thinking, "Okay, another kid with big dreams."

Then he casually drops a number: 135 clients.

I sat up. "You have 135 customers for the SaaS product?"

"No," he corrected me. "135 clients at the agency. But we shut it all down to focus on Talvin. We have 3 customers."

Here is what happened. When Covid hit, everything went digital overnight. Newan and a few mates started building whatever people would pay for—websites, apps, digital transformation. Over a few years, that hustle turned into a 50-person team with clients in 14 countries, working with startups, enterprises, and even governments.

It was impressive growth, until they hit a very boring, very painful wall.

The Bottleneck

They couldn't hire engineers fast enough. Recruitment became the bottleneck choking the business. They started losing big projects simply because they couldn't staff them in time.

So, they did something most "AI startups" only pretend they did.

They built a tool for themselves first.

They built a voice AI recruiter to:

• Talk to candidates 24/7.
• Run proper, structured screening conversations.
• Spit out a ranked shortlist so the human team only spoke to the best fits.

Then comes the important part: They used it internally. On every hire. For almost a year.

There was no Product Hunt launch. No "AI revolution" Twitter thread. Just Talvin quietly running their own hiring process in the background. Only after it had saved their own operations did they ask: "Okay, if this fixed our problem, who else looks like us?"

Where They Are Now

They made the hard decision to pivot. They moved from a service revenue model (safe, unscalable) to a product model (risky, scalable).

While the customer count dropped from 135 to 3, the signal is deafening:

• Solid initial ARR: They have signed enterprise customers including Anytime Fitness, Mindvalley, and Sampath Bank.
• Organic Traction: Ranking #1 for "AI recruiter" via organic SEO.
• Product Depth: A roadmap driven by actual enterprise needs, not hypothetical features.

The 3 Green Flags

As an operator-turned-investor, I look for specific signals that a founder understands the game. Here are the three things about Talvin that fired me up:

1. They Dogfooded It for a Year

Most founders build a product based on a guess or a gap in a market map. Newan built this because his own hair was on fire.

They used Talvin internally for 12+ months. When you use your own product to survive, you build different features than when you build for a pitch deck. You solve the unsexy, jagged edges of the workflow.

2. Real Rollouts, Not Just "Logos"

You see a lot of AI startups with impressive logos on their deck. Dig a little deeper, and you realize they are $50/month pilots that no one uses.

Talvin has signed, rolled-out contracts with massive enterprises. These aren't just badges; they are operational deployments replacing critical workflows.

3. Founder-Led Pragmatism

There is no "Head of Sales" hiding the reality here. Newan is running a relentless, founder-led motion.

He isn't chasing vanity metrics. He isn't burning cash on ads to acquire small, churn-heavy users. He is closing difficult, high-value enterprise deals through sheer pragmatism and product quality.

Why Share This?

I’m sharing these notes for two reasons.

First, I want to start making these advisory conversations public. There are founders out there doing thoughtful, difficult work that never hits your timeline because they are too busy building.

Second, Talvin looks like the kind of company I wish more investors saw early. They are capital efficient, structurally advantaged (Sri Lankan cost base), and validated by real enterprise usage.

If you're an investor, you should have Talvin on your radar.

I’ll be sharing more of these "Advisory Notes" as I meet with founders across the APAC region. If you want to see the next one, make sure you're subscribed.