Drop in a target name. Dealert resolves the entity, classifies it against the TRBC sector tree, identifies its geography and size band, and returns a ranked precedent table. Every match score is decomposed into four visible, editable axes — industry, geography, size and recency — so you can defend the list line by line.
One target (DACH facility services), six precedents shown. Every match score is broken into the four ranking axes so you can see exactly why a deal scored where it did.
Every match score is a weighted sum of four sub-scores. Each sub-score is a discrete function of distance on a tree (industry, geography) or a band scale (size, recency). The weights are visible. You can change them. The same target with the same weights returns the same ranked set, every time.
Type a company name or paste a URL. Dealert finds the canonical entity, confirms the website and looks up the official LinkedIn page. If multiple matches exist, you pick.
An LLM classifies the target against the TRBC L1 → L5 hierarchy and chooses its headquarters country. The classification is shown and editable before you generate comps — you stay in control.
Revenue is estimated from public signals. You can override it. The system maps revenue to a five-band size scale (Micro / Small / Mid / Large / Mega) used by the ranking engine.
Up to 100 precedents are returned, each with a match score and the per-axis breakdown that produced it: industry distance, geography overlap, size proximity, recency.
Generate a ranked precedent set the night before a pitch. Defend the list with per-axis scoring rather than "we picked these because we know them." Drop into the appendix as-is or refine.
Frame an entry multiple against the cleanest precedents — same sector, same region, same size band. Re-weight to test sensitivity. Every number in the memo is traceable to a deal record.
Stop hand-picking comps that flatter the thesis. Show the board the ranked set the engine produced, the weights you used, and the precedents that anchored it. Audit-ready.
The fastest way to judge the engine is to run it on a company whose actual comparable set you already have in your head. Drop in a name and see what comes back. Free account, no card.