SaaStr events run on agentic support — from questions to closed deals

Global events, a three-person team

SaaStr hosts conferences where the biggest operators, investors, and founders in B2B SaaS come to learn, connect, and build.

At SaaStr AI London, that meant 25 sessions, hundreds of attendees, and questions hitting all at once — badge pickup, session recommendations, networking, logistics — with no way to support attendees in person.

“Our team was six time zones away,” said Amelia Lerutte, Chief AI Officer at SaaStr. “We needed something that could answer every question, in real time, without us in the room.”

This audience of founders and executives building with agentic AI expected the event technology to impress as much as the stage. Chatbots had no place here.

Then the challenge scaled. SaaStr Annual: more than 100 sessions, 150 sponsors, 10,000 attendees, and questions around the clock.

An agent for every conversation

Instead of asking three people to answer every question at every event, SaaStr deployed agentic support for each one — Lindsay, powered by Agentforce, for SaaStr AI London, and Amelia AI, powered by Qualified, for SaaStr Annual.

At London, Lindsay drew on each attendee's profile — who they are, what they do, and which SaaStr events they've attended before — to help with session recommendations and agenda building, by chat on the web or by phone through Agentforce Voice.

At SaaStr Annual, Amelia AI works the front door of the event site around the clock. A custom avatar of Chief AI Officer Amelia Lerutte gives the agent a familiar face for a community that already knows her.

Amelia AI also knows who's on the other side of the conversation. When a sponsor arrives to evaluate a six-figure package, the agent sells — asking about goals, suggesting the right package, and booking a meeting directly on a rep's calendar so the rep walks into a warm deal. When a founder shows up to buy a ticket, the agent switches modes — answering questions, offering a discount code, and pointing them straight to registration.

Between the two agents, no one at a SaaStr event waits for help — not the attendee looking for the right session, and not the sponsor ready to spend six figures. The question gets answered, the meeting gets booked, and SaaStr's three people never touch either one.

Grounded data, firm guardrails

For London, Lindsay's setup was deliberately lightweight: session details stored in knowledge articles, attendee context from Salesforce contact records. When testing revealed Lindsay would invent scheduling availability, SaaStr's chief AI officer used Agentforce Builder low-code interface to add a guardrail in natural language — instructing the agent to acknowledge uncertainty instead — no engineering team required.

Amelia AI took that playbook and scaled it. The agent's edge is knowing everything, current as of this morning: it recrawls saastr.com and the event site every day, so when a session moves, the answer moves with it. Content that never touches the website — like the sponsorship prospectus — gets loaded in directly. And for each event, SaaStr gives the agent a smaller, dedicated memory, so a question about a session room searches that event's details, not all of saastr.com.

The agent knows who it’s talking to with customer data. By using SaaStr's unique Agentforce Sales data, Amelia AI recognizes a visitor's company, title, and budget before they type a word. A connection to SaaStr's marketing automation platform adds the engagement history — so the agent knows not just who a visitor is, but how warm of a lead they are.

And roughly 15 guardrails govern what the agent does with all of it — always speak as a SaaStr employee, never quote exact pricing, deflect speaker requests to a formal committee process. When a sponsor is ready to talk, Amelia AI books the meeting directly on a seller's calendar.

That's the build – an agent grounded in what SaaStr knows, aware of who's asking, and explicit about what it won't do.

Inbound that never goes cold

At SaaStr Annual, Amelia AI booked 614 qualified sales meetings at an average ticket size of roughly $85,000. Getting there meant working 2.25 million site sessions and handling 402,000 conversations directly. That's the payoff of an agent that never logs off — every visitor worked, at any hour, in any time zone.

"We're three people. There is no version of this we could have staffed with humans," said Jason Lemkin, CEO and Founder of SaaStr.

Amelia AI also recovers the leads that slip away. A visitor who browses the sponsor page without finishing gets a follow-up with examples of similar sponsors. A ticket buyer who doesn't convert gets a VIP code — a campaign that alone has sold hundreds of thousands of dollars in tickets. 

Every new lead the agent captures flows back into Salesforce – and that’s what keeps the agent getting sharper. When Amelia AI books a meeting, the meeting doesn't just go to whoever's next in line. The agent looks at what each salesperson actually closes — the deal types, the win rates, all sitting in Salesforce — and gives the meeting to the person most likely to win it. Because Amelia AI checks that data every time, the matchmaking gets better as more deals close.

Why SaaStr Chose Salesforce

Native data access

Every conversation Amelia AI has begins with what SaaStr already knows — and ends with Salesforce knowing more. Reads and write-backs happen natively, with no integration to build or maintain. Following Salesforce's acquisition of Qualified, SaaStr is consolidating more of its inbound stack onto the platform. 

Built for more than one audience

SaaStr's site serves two very different buyers — sponsors and ticket buyers — and most inbound tools force one script for both. Qualified let SaaStr encode the distinction from day one, which is why the agent converts instead of just answering.

Guardrails in natural language

Both agents needed explicit rules — Lindsay to never invent scheduling details, Amelia AI to never quote exact pricing. SaaStr wrote those guardrails in plain language, no engineering team required, and tightened them as each event taught the team something new.

Support from Forward-Deployed engineers

For a team building a live event tool with no room for error, hands-on partnership was nonnegotiable. Salesforce’s forward-deployed engineers built Lindsay collaboratively with SaaStr’s chief AI officer — deploying a fully functional concierge in two weeks, before SaaStr AI London’s doors opened.