Competitor dossiers with a confidence on every claim, and a gaps file the system will not guess around.
IN PRODUCTION
This runs in client systems today. The sample below was regenerated for a fictional company.
Two files most companies do not keep. A dossier where every claim carries its source, its confidence, and whether it is fact or inference. And a known-gaps file that says plainly what is not known, who could answer it, and which questions the system will refuse to answer around rather than guess.
TRY IT / RUNS IN YOUR BROWSER
Use it here.
The mechanism is the production one. The company, the catalogue, and every figure are fictional.
Loading the dossiers.
If this message stays on the page, the data at /assets/systems/dossiers-and-gaps/data/ did not load.
THE OUTPUT / GENERATED FOR APEX INSTRUMENTS
What comes out.
JSON
Four competitor dossiers, 58 claims, 18 sourcesData · Calder Metrology the incumbent, Vantage Bench Systems the price challenger, Halden Precision the premium specialist, Orrery Labs the subscription newcomer. Each entry covers positioning, product range, pricing posture, strengths, weaknesses, where they win and where they lose. Every claim names its sources, its confidence and its type. 15 are inferences with the reasoning written out. 6 are low confidence and each one says what is thin about it. Every entry carries a last verified date and the role that verified it.Open the file ↗
JSON
25 known gapsData · Each with the question, the area, why it matters, who could answer it, what it would take, the cost of leaving it open, the date it was opened and whether the system refuses to guess around it. 12 are high impact, 7 have been open more than a year, and 7 carry a refusal line: the exact sentence the assistant says instead of an answer. Two of the gaps were opened by a QA fleet run in the same week and one of them is the question that run raised without answering.Open the file ↗
JSON
15 roadmap items in four honest statesData · 4 committed with an owner and a date, 4 will not do with the reason recorded at length, 4 intended with an owner and no date because nobody has scheduled them, 3 unfunded with what is missing named. Every item lists the gaps it would close. The four refusals are the point of the file: a market sizing study nobody could name a decision for, an estimate from two negotiations, a tracking scheme that would be biased in a direction nobody can size, and a process for manufacturing competitor claims out of job adverts.Open the file ↗
Each claim rests on a source, and the source is named.
02Rate the confidence
Low confidence is recorded as low confidence, not rounded up.
03Separate inference from fact
An inference is labelled, so nobody quotes it as a finding.
04Name the gaps
What is not known, why it matters, and what closing it would take.
05Refuse where it matters
Some gaps are marked as refuse-to-guess, and the system honours that.
WHAT IT READS
Researched claims with their sources
A confidence and a verified date per claim
The open gaps and who could close them
A roadmap with honest states
WHAT IT CHECKS BEFORE IT SHIPS
Every claim carries a confidence and a source
Inference is visibly distinguished from fact
A roadmap without refusals is not accepted
A refuse-to-guess gap is never answered around
In production: competitive dossiers, domain research, and known-gaps and roadmap files in both company systems.
The mechanism runs inside client systems. Everything on this page was regenerated for Apex Instruments, a company that does not exist. Client names, source files, and internal wording stay private.
