The best strategy research already works the way the engine works: individual claims are sourced and graded, contradictions between data points are noted, gaps in the evidence are named, and the final recommendation is honest about its confidence level. What it doesn't do is carry that structure forward to the next engagement, the next client, the next analyst who joins the team.
The structured brief a senior consultant produces for a board is a claim vault — it just doesn't look like one, and it evaporates when the project ends.
Research from engagement A is filed. Engagement E — same sector, adjacent question — starts from zero because nobody has time to find, read, and quality-assess what was done before. The institutional knowledge exists; it is just unqueryable.
The team starts using AI tools. The output sounds authoritative. One analyst spot-checks; most don't have time. A claim that traces to a thin source makes it into the deliverable. The client asks about it in the read-out. There is no audit trail to consult.
A new partner inherits a client relationship. The sector knowledge built over three previous engagements is in six PowerPoint decks and one analyst's memory. The new team rebuilds it in weeks four through seven. The client notices the questions they've answered before.
The client implements the recommendation. Eighteen months later something goes wrong. The question arrives: what was the evidence basis? What was known, what was contested, what gaps were acknowledged? The answer is: we have the final deck, but not the evidence chain.
Verified claims from engagement N are available in engagement N+1. Not as a PDF to be read — as structured, searchable claims with confidence scores, source tiers, and provenance. A new engagement in a sector you've covered before starts with a verified base, not a blank slate.
Every claim in a Epistamate session carries its source tier, the number of providers that corroborated it, whether it survived adversarial challenge, and the formula-computed confidence score. When the client asks "how confident are we in this?", the answer is a number with a breakdown — not a paragraph of hedging.
A new analyst joining a client team can query the verified claim base from prior sessions. They inherit the contradictions that were tracked, the gaps that were named, and the confidence scores that were computed — not just the narrative output. They start informed, not starting over.
When a recommendation is finalised, the Decision Log records the full evidence state at that moment — verified claims, contested claims, acknowledged gaps, confidence scores. This travels with the engagement record. Eighteen months later, the evidence basis is not in the analyst's memory. It is in the log.
Smaller teams with deep sector knowledge and no dedicated knowledge management infrastructure. The claim graph is the knowledge management system — it compounds with each engagement without requiring a separate KM investment.
DD is structured evidence work at its core — claims about market size, competitive dynamics, management track record, regulatory risk. These need provenance, confidence gradations, and audit trail. The engine does all three as a byproduct of how it works.
Solo practitioners and small expert practices doing sector-specific advisory work accumulate institutional knowledge with no infrastructure to preserve it. The knowledge graph turns that accumulated expertise into a queryable asset rather than a perishable one.
We're talking to consulting practices and advisory firms in active development. We'd like to understand the specific workflow problem before describing the solution.