The Leverage Discovery

A four-person agency founder was certain that proposals were his bottleneck. Eight hours of writing per opportunity, a third of them going nowhere. He had asked for help with the proposal itself — a template, better prompts, a generator.

Discovery went differently. We walked through the last three proposals in detail, end to end, including the conversations that preceded them. The pattern appeared in the third example: the proposals were well-written. The qualification before them was not. Half the hours went into proposals that should not have been written — for prospects whose budget, fit, or decision process had never been clarified.

The leverage was one step upstream of the fire. We designed a ninety-minute qualification conversation, running before any proposal work began, with five explicit questions the founder now asks before touching a document. The proposal-writing itself was unchanged.

A solo consultant with twelve active opportunities arrived with a question about which CRM to adopt. He had tried three. None had stuck. He wanted help picking.

The walkthrough revealed the real problem within ten minutes. The tool question was a surface; underneath, the status of each opportunity lived nowhere except in his working memory. He could not keep a CRM updated because the CRM required writing down what he already knew — a task that felt pointless in the moment and expensive in aggregate.

The leverage was not tool-selection. It was the moment of capture. We designed a fourteen-field end-of-call note, completed in the last five minutes of every conversation, readable in ten seconds. The notes themselves became the CRM — no separate system, no double-entry, no ceremony.

A strategist ran weekly research sessions that produced, she estimated, about one-fifth of their intended output. The request was for better prompts — ways to make the AI assistant produce cleaner summaries.

Discovery looked at a single session, minute by minute. Of ninety minutes, seventy-two went to orienting — re-reading prior notes, reconstructing what had already been decided, getting back into context. Eighteen minutes went to actual finding. The prompts were fine. The ratio was broken.

The leverage was splitting the session into two different kinds of work — each with its own shape. Context preparation became a morning task, written out beforehand as a short brief. The research session itself started already oriented, went immediately into synthesis, and ended with a verified claim. Output tripled. Time halved.