Supplied evidence
Documents provided for the screen are identified as sources rather than treated as an undifferentiated data room.
Aquerim methodology
Evidence first. Inference labelled. Gaps left visible.
The methodology is designed for a screening decision: establish what the early-stage file supports, reconstruct a defensible earnings view, surface risks and direct the right questions to professional review.
Evidence hierarchy
Four layers that should not be collapsed into one answer.
Public-record context
Where relevant, public information can be compared with the supplied position and any mismatch stays visible.
Analytical inference
Interpretation is labelled as judgement. It is not presented as a verified fact merely because a model produced it.
Unresolved question
Gaps, contradictions and unsupported claims become explicit items for the buyer or adviser to test next.
Defensible earnings
A bridge the buyer can inspect, not a single magic number.
The screen separates reported profit, owner remuneration, seller adjustments and the adjustments that can be supported by the material reviewed. The result is an earnings view with its limitations beside it.
The synthetic Copper Kettle example starts from £119,000 of presented earnings and reconstructs £26,100 of defensible earnings, making the resulting 11.3× asking-price multiple visible for challenge.
Where AI is used
AI assists the review. It does not turn uncertainty into fact.
AI assists document extraction, comparison and synthesis across the early-stage file. The workflow is designed to keep source evidence, model-assisted interpretation and unresolved gaps distinct.
Quality gates and a separate QA challenge test contradictions, completeness and calibration before the report is rendered. Findings still require independent verification by the appropriate advisers.
Test the method against the output