From your question to a named decision in 24 hours.
The short version: AI compiles public-source research; Damyan vets every source, synthesises the evidence, and writes the recommendation. Every factual claim is traceable. Every report ends with a named next move — or a refund and a written explanation of why the evidence won’t support one.
The long version is below.
From brief to named next move.
24 hours. Every step accounted for.
- 01YOUHOUR 0 · 5 MIN
Send the brief.
One paragraph. The decision in scope, the company, the question. No deck. No call. The form is four fields.
- 02AIHOUR 1 · ~1H
AI drafts from primary sources.
Filings, archived pages, public datasets. Faster than any manual analyst on the compilation pass. Citations attached to every claim.
- 03DAMYANHOURS 2–20 · ~18H
Human reviews end-to-end.
Every claim re-verified. Sources cross-checked. The judgement of which evidence counts is the part you're paying for.
- 04INBOXHOUR 24 · 09:00 CET
Named next move on the cover.
PDF + working files + Loom. The specific decision the evidence supports. If the evidence won't support one, we say so on the cover and refund.
AI compiles the research. The judgement is human.
AI is faster than any analyst at compiling public-source research. It is not reliable at picking which sources to trust, what the evidence implies, or which decision the data supports. That part is what you’re paying for. Here’s the split, line by line.
| AI DOESResearch compilation | DAMYAN DOESSource vetting · synthesis · recommendation |
|---|---|
| Source discovery — surfacing filings, public datasets, news, archived pages | Source selection — deciding which of the surfaced sources are credible enough to cite |
| Draft synthesis — initial competitor profiles, market summaries, financial pulls | Claim rejection — cutting any claim that doesn't survive a sceptical-CFO read |
| Initial tables — pricing comparisons, segment sizes, channel CAC estimates | Commercial judgement — what the table means for THIS business, in THIS quarter |
| First-pass cross-checks — flagging contradictions between sources | Final recommendation — the named next move, the specific number, the deadline |
| Formatting & ToC, citations index, working-file generation | Cover-page write — the deliverable the buyer pays for |
The gate:no report ships until each claim passes a source / relevance / decision-usefulness check. If a claim fails any of the three, it gets cut. If the cover-page named-next-move can’t be supported by the surviving evidence, the brief is refunded and we say so on the cover. Read the full methodology →
The three checks every claim has to pass
- Source. Is this from a primary or otherwise credible source (filing, dataset, archived page, transcript) — or just a downstream summary? Downstream-only claims get cut.
- Relevance. Does this claim actually bear on the decision in the brief, or is it interesting-but-unused context? Unused context gets cut.
- Decision-usefulness. If the named next move on the cover page is contradicted by this claim, one of them changes. If both can stand, the claim earns its place.
If the surviving evidence won’t support a named next move, the cover page says so and the brief is refunded. A hedged report isn’t worth shipping.
What you receive
- PDF report — ~30 pages on Starter, 18 pages on Deep-Dive. Named next move on the cover. Every factual claim cited.
- Source log — every URL, filing, dataset, and transcript referenced, with the access date.
- Working spreadsheet — the underlying numbers behind any tables in the report (where applicable).
- Loom walkthrough — a 5–10 minute screen-recorded walkthrough of the key findings.
- Async written replies — reply to the delivery email; written answer back within one business day. No calls.