Backstory's 'Situation Search' Puts Deal History Inside the AI Tools Reps Already Use

People.ai, now operating under the Backstory brand, built a way for reps to ask “how have we handled this before?” directly inside Claude, Copilot, or ChatGPT — and get back real comparable deals with their actual outcomes.

The capability, called Situation Search, is delivered through an MCP server and addresses a specific failure mode: when a deal stalls over a procurement delay or legal pushback, reps have historically had to track down whichever manager or senior colleague happened to remember a similar situation. With Situation Search, a rep can ask directly and get matched comparable past deals with real outcomes, without waiting on a particular person’s availability or memory. Backstory positions it as capturing institutional knowledge that would otherwise leave when experienced reps do, standardizing top performers’ winning moves across an entire team rather than leaving that expertise siloed with a handful of individuals.

This is exactly the shift Forrester’s David Mooter flagged a week earlier: he argued agentic AI’s real bottleneck isn’t more model capability, it’s integration — an agent without integration is “merely an answer engine” that can’t take action, and organizations with high agentic-AI readiness are prioritizing integration and change management, not model capability, over the next 12 months. Situation Search is the sales-specific version of that argument: an MCP server is the integration layer turning historical deal data into something queryable inside a general assistant rather than a standalone tool. It also lands on ground Anthropic’s own Cowork usage data flagged as underserved: sales and revenue operations made up just 4% of Cowork sessions in Anthropic’s May sample, even as “business process and operations” work — the category Situation Search targets — was the single largest at 33.4%.

The reps who benefit are the ones who ask the sharper question — not “what happened with this account” but “how did we get out of a procurement delay when this deal looked like it was dying” — because the tool only returns what it’s asked, and a vague query gets a vague comparable back.