Eden Gives Revenue Teams an AI Buyer to Drill Against – Built From Their Own Product and ICP
Sales reps learn to sell by doing. The problem is that most of the doing happens on live prospects, where a blown discovery question or a fumbled objection costs real pipeline. Eden, a Y Combinator Summer 2025 startup, is building the practice layer that comes before the live call — an AI buyer simulator that lets reps run discovery, demo, and negotiation drills against a realistic AI buyer built from their actual product and ICP.
Eden’s core product is simple in concept and sharp in execution. A sales rep or manager pastes a product description and ICP into Eden, picks the type of call they want to drill – discovery, demo, or negotiation – and Eden generates a live AI buyer with their own pain points, objections, personality, and tells. The buyer pushes back in character, stalls when a question lands wrong, and can end the call if the rep blows it badly enough. After the session, Eden returns a full transcript and a scorecard grading the rep against the goal, with every point backed by a timestamp from the call. The result is AI sales roleplay training that reflects the actual pushback a rep will hear in the wild – not a templated script with swapped names.
The company was founded in 2025 by Alex Talamonti and Jason He. Talamonti, co-founder and CEO, was previously a software engineer at Datadog, where he worked on the data visualization and analysis platform used by some of the world’s most demanding customers. He also led operations at Phia and is a graduate of NYU Stern’s inaugural BTE program, where he was recognized on Poets & Quants’ list of notable undergraduate business students. He refers to himself, with some good humor, as the person who got tired of watching reps lose winnable deals because they never got enough reps in. Jason He rounds out the two-person team with deep technical instincts and a shared conviction that sales training has been broken for too long.
The market problem Eden is solving is well-known inside any revenue organization. New reps spend months ramping before they can handle a tough discovery call without a manager on the line. Tenured reps drift into bad habits between coaching cycles. And the conventional fix – call shadowing, role-play with managers, Gong review – is time-intensive, inconsistent, and hard to scale. Eden’s AI sales roleplay training compresses that loop: a rep can run five discovery drills before breakfast, each against a different buyer persona, and come to their next live call with muscle memory that didn’t exist the week before.
The product integrates with the tools sales teams already live in – Salesforce, HubSpot, Gong, Chorus, Salesloft, Slack, and Microsoft Teams among them – so buyer generation can pull from real ICP data and deal context rather than from a blank prompt. Managers can review their reps’ sessions, read the transcripts, and comment on specific moments in the call. For onboarding, that means new hires get reps before they touch live pipeline. For tenured sellers, it means drilling the specific call that’s coming up – a tough renewal, a competitive demo, a negotiation where procurement is pushing hard on price.
Eden is early and the team is small, but the thesis is unusually clear. The gap in sales enablement has never been content – it has been practice. AI sales roleplay training that generates a buyer from your actual product and ICP, grades the call against a real goal, and returns evidence-backed feedback is the kind of tool that pays for itself the first time a rep walks into a hard call prepared instead of winging it.