Building a One-Page Pre-Call Dossier With AI Research

Recent news or public statements relevant to the account, who else is likely involved in the deal based on org structure, the tech stack currently in place, a relevant case study or reference point, and a specific pain hypothesis for this particular call. Anything beyond that is detail a rep can look up later if the call actually calls for it.

A dossier that tries to cover everything ends up being read by nobody. A short one built around what's actually relevant to this specific call gets used every time.

A one-page dossier usually covers these items:

• Recent news or public statements relevant to the account, kept to what bears on this specific call.
• Who else is likely involved in the deal, inferred from the account's org structure.
• The tech stack the account currently has in place, so the rep knows what you would sit beside or replace.
• A relevant case study or reference point the rep can mention naturally during the conversation.
• A specific pain hypothesis for this particular call, stated so the rep can test it in discovery.

Pulling and summarizing public information, recent press, filings, LinkedIn activity, that a rep could technically find manually but shouldn't have to spend an hour assembling before every single call, is where this genuinely helps. The research itself isn't hard, it's just slow when done by hand across a dozen calls a week.

This is the same kind of structured, checkable task that's safe to automate broadly, since the underlying sources are public and verifiable.

AI-generated summaries can attribute a stale headline to a company that already resolved the issue, or state a specific detail with more confidence than the underlying source actually supports. A rep should verify anything they plan to say out loud in the room, especially a specific figure or a claim about a recent event.

Treat the dossier as a strong first draft, not a finished script, and spend the last minute before a call double-checking anything you're planning to reference directly.

Here is a simple rule for the last minute before a call: anything the rep will say out loud gets checked against its source, and anything used only as background can stay unchecked. If the dossier says a company recently suffered an outage, confirm the date and whether it was resolved before mentioning it, since a stale headline sours a call fast. Spend the checking time on names, figures and recent events, and skip the rest. A short habit like this costs far less than opening with a claim the buyer immediately corrects. Afterward, note in the shared log any item the AI got wrong.

A rep preparing for a call with a mid-market logistics company gets a generated dossier noting a recent executive hire, the company's current warehouse-management platform, and a relevant case study from a similar-sized customer. The rep confirms the executive hire is accurate through a quick LinkedIn check, since that's the detail they plan to open with.

The generated pain hypothesis, about integration friction between two specific systems, turns out to be slightly off once the rep asks about it directly on the call, but having a specific starting hypothesis, even an imperfect one, made the discovery conversation sharper than starting from a blank page.

One page, scannable in about two minutes, with the most call-relevant detail at the top. A ten-page report full of exhaustive background research looks thorough but rarely gets read in full before a call that's often scheduled back to back with three others that day.

If a rep has to scroll past several screens to find the one detail that matters for this specific call, the dossier has failed at its actual job regardless of how much research went into producing it.

Use one shared template across the team so every rep knows what to expect and where to look for a given piece of information. Build a habit of flagging when the AI got something wrong, out loud in a team channel or a shared log, so the prompt or process improves over time instead of the same kind of mistake recurring quietly across every rep's calls.

Reps who walk into a call genuinely prepared tend to outperform the flat 19% average new-logo win rate¹ by a meaningful margin, which is the real case for investing in this at all rather than treating it as a nice-to-have.