SaaStr runs an agentic business on one foundation with Salesforce’s Headless Toolkit

Twenty agents, by design

SaaStr runs the world's largest community for SaaS founders including events, content, and a venture fund. They do it all with three full-time employees and more than 20 AI agents.

That agent count is deliberate. SaaStr's community is built on knowing the tools founders actually use, so the team runs best-of-breed agents across sales, marketing, and event operations, each one excellent at its job.

The ambition was making them more than the sum of their parts. Every outreach email, every event conversation, every engagement signal an agent picks up is knowledge the next agent could use but every agent needed to see it first. Twenty specialists were doing twenty jobs. SaaStr wanted them working from one brain.

Nine years, one foundation

To build into their foundation without replacing existing investments, SaaStr standardized on Sales Cloud as their system of record, bringing nine years’ worth of customer interactions, event registrations, engagement signals, and account records into a single source of truth.

“When we deploy new agents in Go-To-Market, we integrate them with Salesforce as the backbone of our central data repository,” said Lerutte.

As more of SaaStr’s tools connect to Salesforce, every customer profile gets more complete — past interactions, engagement signals, and event history unified in one place. The more accurate the profile, the more personalized every agent’s outreach becomes.

API-first architecture

Storing customer data in one place was only half the answer. SaaStr also needed every agent — including the ones built outside Salesforce — to reach that data without a custom integration for each. That's where Salesforce’s Headless Toolkit comes in. An API-first architecture makes Sales Cloud data and customer context accessible to Agentforce and external AI agents.

Agentforce Sales serves as the shared memory layer across nearly 20 AI agents, while Headless Toolkit provides the API fabric that gives Agentforce, Slack, Replit, and other tools access to that context.

With that setup, agents can run workflows across systems using the same continuously updated customer context. One agent enriches a lead, another triggers outreach, another summarizes engagement history — all against the same customer profile, without requiring custom point-to-point coordination between tools.

Our AI agents work directly at the API level," said Jason Lemkin, CEO and Founder. "I had never logged into Salesforce before we deployed the Headless Toolkit. Now I can talk directly to our agents, and they know everything Salesforce knows."

As agents complete tasks, outputs and interaction history are written back into Salesforce, creating persistent context across the AI stack.

Richer, more effective interactions

SaaStr sees the impact of their shared context layer most clearly in Hexi, their Agentforce powered SDR agent.

Rather than sending generic follow-ups, Hexi accesses customer context through the Headless Toolkit, drawing on nine years of CRM history stored in Agentforce Sales: who the prospect is, which events they attended, what they engaged with previously, and where the relationship last paused.

Hexi’s outreach reflects the cumulative work of every agent connected to Salesforce. As other agents research accounts, update records, summarize conversations, and track engagement across systems, their outputs flow back into Salesforce and become immediately available to Hexi.

Running outreach campaigns to more than 3,000 previously unresponsive leads, Hexi achieved a 72% open rate — the highest across SaaStr’s AI stack — because it operates from a complete, continuously updated view of the customer rather than fragmented data spread across disconnected tools.

“Pick one source of truth to store data and build context for your agents. We use Salesforce,” said Lerutte. “When you do that, agents know everything your Salesforce knows, and you can use that context to guide the actions they take.”

Next, SaaStr plans to extend that architecture with a unified orchestration layer in Slack, allowing agents to coordinate actions in real time while continuously feeding activity back into Salesforce.

Why SaaStr chose Salesforce

Unified customer profiles 

Because every agent reads from the same Salesforce foundation, every customer profile reflects the full picture — past interactions, event attendance, engagement signals. Agents work from everything SaaStr knows, not a fragment of it — and every interaction makes the next one stronger.

Open API architecture

Built for API-first operations, the Headless Toolkit gives nearly 20 tools shared access to customer context — no custom point-to-point integrations, and no new connector each time an agent joins the stack.

Native CRM integration

Every other AI tool SaaStr ran required a custom connection to access CRM data. Agentforce read it natively from day one — and every agent since has inherited that head start. Adding a new tool no longer means building a new bridge.