AI Sales Agents

What we talked about.

In this episode of the Sunny Ray Show, host Sunny Ray talks with Nicolas (Nico), co-founder of Aithon, a go-to-market platform built for companies selling into financial services with a deep relationship with AWS. Nico spent ten years on trading floors in financial services before spending another ten years at AWS, and he explains how the discipline and data rigor from trading carried directly into how he builds software today. He describes the moment he realized how broken pipeline data really was, when only 0.74 percent of a customer's emails were linked to any deal in their CRM, leaving over 99 percent of communication data unused. That insight led Nico and his co-founder Nitin to build a deal association engine that pushed that number to 12 percent, a sixteen times improvement. Nico also shares candid lessons from Aithon's rocky first six months, why he left AWS in January 2025 to start his third career chapter, and how Aithon turns raw CRM data into leads, alerts, and coaching for sales teams.

A former AWS and Wall Street trading floor veteran explains how Aithon turns messy CRM data into real pipeline intelligence.

The questions, and the answers.

What are you building at Aithon?

Aithon is a go-to-market platform built for companies selling into financial services, with a strong relationship with AWS that helps customers manage that partnership too. My co-founder and I have each spent ten years at AWS, so we saw firsthand how the sales tech stack was or wasn't working. We've been building Aithon for about eighteen months, using that experience as inspiration for a sales stack we believe can actually accelerate sales.

Ten years at AWS and before that trading floor technology inside banks, what carried over from the trading floor into how you build today?

On the trading floor I supported everything from high frequency to fundamental desks, and it taught me operational and technical discipline matters because you're sitting so close to the money. Regulation also demands rigor, which carries over to our financial services customers today. The other big lesson was data quality. If you have the wrong ticker or price, your trade is worthless. It's garbage in, garbage out, so we obsess over making sure our data is sharp and well connected.

When did you first notice that nobody could answer basic questions about their own pipeline?

Working with sellers across geographies at AWS, I saw two types: some stayed strategic on every detail, others handled day to day tasks but never revisited planning after Q1. There's constant background noise at a company that size. Then at Aithon, when we integrated a customer's CRM, we found only 0.74 percent of their emails were associated with any deal, meaning over 99 percent of their communication data was unused, a terrible foundation for answering basic pipeline questions.

What made January 2025 the month you walked away from a ten-year run to start this?

I loved my time at AWS, the culture and the people, but sometimes you have to pull yourself out of a situation to do something new. After ten years in trading and ten years at AWS, I felt ready for a third chapter, and entrepreneurship was the right move given my finances, family, and readiness. Once I saw this problem, AI had finally made human driven sales conversations analyzable, so it felt solvable, and my co-founder Nitin and I found each other at the right time.

What did you do when you first saw that only a fraction of a percent of emails were attached to the right opportunity?

The first thing I did was dive deep into the data, which I learned from my AWS days. Every sales leader knows sellers hate logging data into Salesforce or HubSpot, so it wasn't surprising, just a bit shocking. I realized it wasn't a hygiene problem you fix by telling sellers to tag better. Instead we built clever heuristic and AI driven reconciliation, what we call our deal association engine, taking email association from 0.74 percent up to 12 percent, a sixteen times improvement.

You call the deal record work the unglamorous half. Why is that unglamorous half the actual moat?

Moving data around is fun for a technologist, but it's not what brings value to users, so I'd call it unglamorous from that standpoint. The glamorous part is the insight it enables: next best actions on a deal, alerts when a deal goes dormant or a stakeholder loses patience, relevant news on an account, and now an agentic platform where one of our most fun agents finds underused AWS credits and marketing funds that partners are entitled to.

What did you guys get wrong in Aithon's first six months?

Pretty much everything. This was my first venture and my co-founder's second, and our first roadmap had so much on it that I scared the team by saying we had three months to build it all. I was unrealistic. I'm also good at hiring but not always fast enough making tough calls when someone isn't working out. And relying too heavily on any one person is a real risk at a startup with just a handful of people.

Can you explain Aithon to a revenue leader who's convinced their CRM is basically fine?

It depends on your expectations. If your CRM is just a system of record for pipeline status, it's probably fine, but most sales organizations actually work for their CRM instead of the other way around. At Aithon we combine private data like emails and calls with public data like news, LinkedIn posts, and filings to generate qualified leads, draft human sounding outreach, and raise alerts when deals stall, turning the CRM into a strategic advantage instead of just a record.