Clay’s Sculpt conference attracted 2,000 attendees from 30 countries to Pier 48 in San Francisco, and you could feel the energy in the room.
FastForward CRO Chaz Gerads and I spent Partner Day and Sculpt with the incredible Clay team. As a Clay certified partner, FastForward gets to see where the product is headed before our clients need it, and those two days were some of the most valuable of our year.
Here’s what we learned about where Clay is headed, and what I think it means for the CEOs and revenue leaders we work with.
Clay is evolving into a self-learning revenue engine
Clay always talks about GTM alpha: using the unique tools, systems, and custom plays you build in Clay to gain an edge on your competition. Every edge has a shelf life, so the teams that win are the ones who capture learnings from their work and keep improving. In Clay’s words, “there’s no permanent competitive advantage in GTM, only the continuous pursuit of temporary ones.”
From 2018 to 2022, the edge came from volume. Load up a sales engagement platform and let it rip. That still has its place, but in 2026 almost anyone can launch a campaign. Clay made the same point on stage: at last year’s Sculpt, the bottleneck was execution capacity. A year later, the real edge comes from “how well they can learn.”
This year, Clay’s Co-founder and CEO Kareem Amin shared what Clay is building to make the learnings about your competitive advantages automatic: a GTM system where agents work your full TAM, learn what converts, and act on your behalf. Teams get smarter every time they run a Clay campaign or play. This is a huge next step in accelerating GTM alpha as you build.
The part that stuck with me is how much evidence most teams leave on the table. Clay points out that on any given day a company has thousands of emails going out, signals firing on hundreds of accounts, and sellers holding dozens of conversations. Until now, teams could only learn from the small share they had time to review on a quarterly basis. Agents can read all of it. Most revenue teams still learn at the speed of the QBR, so that is a massive shift.
Kareem summed it up well in his post after the keynote: “One agent helps your GTM engineers build. The other helps your reps sell. Both turn results into recommendations in Clay, so a sales conversation can shape your next campaign.”
The three building blocks
Clay is building three core elements to make this a reality.
1. Workflows
One canvas for an entire GTM play, with branching, code, and agents. Plays trigger when an audience segment updates, so they always run on fresh data, and every record is traced so you can see exactly what happened and re-run the fix.
Clay describes Workflows as one visual space where “every step can branch, loop, run code, or call an agent,” running across millions of records. You can build a workflow by describing the play in plain language, either in Clay or through the coding agent of your choice via CLI. Workflows can trigger when records enter or change in a segment, then write results back to Audiences, so the context Clay’s agents reason over gets richer with every run.
The tracing is what I’d point a revenue leader to first. You can follow the exact path a record took, including the reasoning applied at each step, then fix the play and re-run it in bulk. Workflows is in open beta for all Clay users today.
2. Clay’s Global Agent + Knowledge Hub
Give Clay a goal like “50 more opportunities this month” and the agent builds and runs the play, checks results, and suggests ICP updates. Knowledge Hub is a living record of your business context that the agent reads and writes to, so every future play inherits the lesson. You approve high-stakes actions before they run, and every credit traces back to the run that produced it.
The keynote demo showed what that loop looks like. An agent ran an account-based campaign for a fictional payments company and booked six meetings from 18 replies within a few days. It noticed four of the six were multi-site healthcare operators and suggested adding that segment to the ICP. Once a person accepted it, every agent working on the account used the updated ICP.
On control, Clay says every action comes with an approval tier, a cost, and an artifact you can point to, and Clay asks before anything consequential happens. The global agent went live for Sculpt attendees on paid plans, with early access rolling out to more users.
3. Inbound SDR Agent
Qualifies leads by voice and gives a live, tailored demo of your actual product to the leads reps would leave on the table. It learns from your best sellers’ call recordings and the ICP and business context already in Clay.
The Inbound SDR Agent asks qualifying questions by voice, updates lead scores in real time, and then books the meeting with the right rep or demos your product in a sandbox. Kareem’s example: someone requests a demo at 9pm on a Friday. The agent answers their questions, demos the product, and qualifies them in one conversation. Clay says its conversations also feed recommendations on ICP, positioning, and competitive differentiation across every campaign.
The CRO in me loves this one. Clay is clear that it isn’t built to replace your SDR team. It’s built to catch the lower-score leads your team doesn’t have capacity to work, the “diamond-in-the-rough prospects” who would otherwise get disqualified. It’s in preview now, with a waitlist and a phased rollout starting this fall.
What it means for GTM leaders
Agents are now the critical baseline for every GTM platform.
GTM alpha depends on the context agents learn from: a written ICP, sharp positioning, clean CRM data, customer call and demo recordings, and governance to act safely on live records. Clay says it plainly: a self-learning system needs the team’s ICP and positioning “written down — not just in slide decks and people’s heads. An agent can only improve a strategy it can read.”
That’s where most teams will feel it. In the AI era, we need to get GTM teams to work differently. It isn’t creating new processes from scratch. It is unwinding years of outdated, slow, manual processes and getting people to reorient themselves on how to do it better. It is a lot harder than it sounds. An agent pointed at a messy CRM and an ICP that lives in three people’s heads will learn the wrong lessons, faster.
Clay is building guardrails into the product with approvals and credit tracing. Each team still decides where those guardrails go. Which actions need a human? What’s the credit budget for a play? Who signs off when the agent suggests changing your ICP? Those are leadership decisions, not settings.
Clay says GTM engineers will manage agents, the same way many software engineers now manage agents that write code, and that reps will spend more of their time on the parts of selling that rest on trust. I agree, but someone still has to build the foundation those agents stand on.
Where FastForward fits
That foundation is our daily work with B2B revenue teams. We assess, build, and deploy new systems for GTM teams, and the agent era makes that work matter more, not less.
We turn tribal knowledge into context an agent can use: the ICP that lives in your best rep’s head, the positioning buried in a pitch deck, and the lessons sitting in call recordings nobody has time to review. We architect Workflows across our clients’ full TAM, and we set up credit budgets, observability, and human approval so leaders trust the engine. Then we help retire the legacy stack piece by piece.
We don’t do it all at once. Start with one play, prove it, and build from there. Brick by brick.
The teams that win here will be the ones who did the foundational work first.