What SaaStr 2026 taught us about the state of GTM AI

We came to SaaStr Annual 2026 expecting a lot of noise about agents. We got that. But underneath the noise, there were three conversations worth paying attention to – and one honest question that kept resurfacing when the demos were done and the real talk started.

Here’s what we took away.

1. Agents are everywhere. The results, not so much.

If 2025 was the year everyone started talking about agentic AI, 2026 is the year everyone arrived at SaaStr having already tried it – and most of them with a mixed report card.

The shift in the room was striking. A year ago, the question was should we be looking at agents? This year, the question was why aren’t ours working the way we expected? The technology had moved fast. The implementation reality had moved slower.

The gap between time invested and outcomes delivered is real, and it was one of the most honest recurring themes of the conference. Deploying agents is one thing. Getting them to reliably work for you – to actually close the execution gap rather than create new ones – is another thing entirely.

“Agentic” has officially become a word that means everything and nothing. The interesting question is no longer whether you use agents. It’s what you ask them to do, and what guardrails you’ve built around them.

2. Data quality is the unsexy blocker nobody wants to admit

Every good demo at a conference looks clean. The data is tidy, the CRM is current, the outputs are sharp. Production is a different story – and at SaaStr this year, more people were willing to say so out loud.

Data quality is still the blocker nobody wants to talk about in a pitch, but everyone complains about the moment they go live. And for GTM teams specifically, the problem runs deep: first-party data alone isn’t enough. Your CRM and conversational data tell you what happened. They don’t tell you who else looks like your best customers, what intent signals are firing right now, or what firmographic context would change how you approach an account.

For AI to work in a revenue workflow, it needs a rich foundational data layer – first-party combined with third-party sources like Apollo, intent data, and firmographics. Without that, you’re not running intelligent agents. You’re running confident ones. And confidence without context is a liability.

Spray and pray is dead. Everyone’s inbox is full. The only way to cut through is to get genuinely granular about who to reach out to and with what message – and that starts with getting the data layer right, before anything else.

3. The best answer right now? Forward Deployed Engineers

If there was one consistent answer to the question of how do you actually make agents work in production, it was this: someone who sits at the intersection of technical depth and customer context.

Forward Deployed Engineers = FDEs, came up repeatedly as the practical bridge between the promise of agentic AI and the reality of deploying it inside a live GTM motion. Not engineers who throw things over the wall. Not consultants who arrive with a framework. People who understand the technical architecture and the commercial reality, and can hold both at once.

We work with FDEs at Airspeed, and our customers are increasingly seeing the value too. When you’re trying to close the gap between what an agent can theoretically do and what it actually does reliably for a specific team with specific workflows – that human judgment is the difference.

What we took into Wednesday

These conversations didn’t stop at the conference floor. On Wednesday, we co-hosted the Gradient Descending roundtable with Samuel Colvin, co-founder and CEO of Pydantic, to go deeper on agent engineering – specifically forward deployed engineering, and how you actually build reliable, production-grade agents that teams can immediately benefit from.

The thread running through all of it: the companies pulling ahead right now aren’t the ones with the most agents. They’re the ones who’ve been most honest about what agents need to actually work – clean data, clear guardrails, and people who can build at the intersection of technical and commercial.

A note on what to actually look for

For anyone actively evaluating GTM solutions right now, a few things worth pressure-testing before you sign:

Don’t just ask whether they offer agents. Ask whether they’re truly AI-native, or just wrapping old workflows in a new language. The distinction matters enormously in production.

Dig into the data layer. The quality of the output is only as good as the data going in. Ask what first-party and third-party sources they’re drawing from, and how they combine them.

Check what value they add on top of your existing tools. Watch out for platforms that promise a lot, then leave you exporting data into another AI tool just to get something useful. That’s two steps for something that should happen in one – or better, automatically.

Intelligent agents, paired with human relationship-building, is the answer. Agents alone aren’t.

It was a great few days. The honest conversations, the ones about what’s actually hard, not just what’s theoretically possible, were the most useful ones we had. We’re looking forward to continuing them.