How a Deal Intelligence report is built

Deal Intelligence methodology

AI-driven monitoring of the public record, editorial selection, document-first research, and a stable page architecture with edited headings.

Published 16 Sep 2026 · Updated 22 Sep 2026 · Method changes only

Deal Intelligence is not investment, legal, tax or accounting advice. Every material fact is taken from documents that are already public, or from arithmetic on those documents. Read the independence and disclaimer before you treat a number as something you can act on.

A Deal Intelligence report starts with a public announcement and ends with a page a practitioner can use. The work in between is selection, document gathering, and editorial judgment. This page is that pipeline, written so a reader can see why one deal ran and another did not.

Acquiry uses AI-driven monitoring to watch public announcements as they land: issuer investor-relations pages, exchange filings, regulator notices, and named newswires that circulate those documents. The system clusters duplicate headlines that are the same buyer and target told five times. It does not scrape paywalled article bodies, and it does not treat a blog rumour as an announcement.

The scan is a watchlist, not a verdict. A cluster still has to clear an editorial bar before anyone writes a report.

Headline lane

The headline lane is for the day's most material public M&A, anywhere. Size helps. A complete public record helps more. A $20 billion take-private with a merger agreement and an 8-K will beat a noisy mid-market rumour with no document. Media corroboration (more than one serious outlet circulating the same filing) is a usefulness signal, not a substitute for the filing.

Gaming lane

The gaming lane is narrower on sector and wider on size. Casino, iGaming, betting, lottery, sportsbook, or a games company where the deal is about that business. Any ticket size, including a small cash take-private, if it is a company transaction and the public record can support a report.

Gaming skips, on purpose:

• Media or sports rights with no equity in an operator changing hands.
• Property, hotels, or real estate that happen to sit next to a casino.
• A licence grant or renewal with no acquisition.
• Talent, customer, or "user acquisition" headlines that are not M&A.

A Flutter-scale deal can qualify for both lanes. It still publishes once.

Material, for this desk, is a mix of:

• Control or a clearly disclosed stake (not a vague "strategic investment" with no number).
• A document trail a reader can open.
• An operating thesis that can be explained: capability, distribution, regulation, product.
• Enough public facts that the article will not become a catalogue of what the issuer refused to say.

Undisclosed price is allowed. Entirely undisclosed economics plus an empty operating story is not. Opaque headline deals can still run when the public seller or buyer record is rich enough to explain the combination. Inventing a purchase multiple to fill the hole is not allowed.

Once a deal is chosen, the desk pulls the closest primary documents: the announcement, any merger agreement or scheme booklet the issuer filed, the last audited accounts, the relevant licence or regulator page, product documentation the companies already publish. Secondary press is used to see how the announcement travelled and what competitors said in public, not as a replacement for the filing.

The internal file is larger than the article. That is deliberate. Research depth creates authority. The published page is a selection. Every paragraph has to improve a reader's understanding of the deal, the companies, the product, or the industry. Facts that do not do that stay in the file.

The architecture is stable so a returning reader knows where to look. The headings are not.

• H1 and title match, under 110 characters, entity-led, no "seismic" or "game-changing".
• Standfirstis one specific sentence, not a restatement of missing terms.
• News leadis inverted-pyramid fact, before any deal-summary grid.
• Deal summaryis an HTML table of parties, consideration if public, status, jurisdiction, close window if given.
• Analysis sectionsfollow the deal: what was bought, why it fits, how the stack works, where it can be distributed, what it says about the market, how combination could work. Every report carries at least one What this means aside under the densest mechanic. They are not a new chapter and not a dumbed-down twin of the article.
• Acquiry viewsynthesises. It does not recap the piece.
• FAQis deal-specific. The dek names what the questions actually cover.
• Source ledgeris the bibliography. Related reading points at other Deal Intelligence reports, not at the press release the ledger already lists.

Each report carries original diagrams in a single navy / cyan / gold language. They are topologies of the argument in that section: holding structures, capability stacks, routing, combination. They are not stock photography, not people, and not consumer gambling tropes. A figure is hidden from assistive tech only when a sibling data table carries the same structure for crawlers and screen readers. Images lazy-load. Alt text describes the transaction fact the picture is illustrating.

The desk is built for weekday publication. Weekends are quiet unless a genuinely outsized public deal appears. If monitoring finds only rights deals and property headlines, gaming is skipped and the skip is recorded internally. The public site does not invent a "today in iGaming" post to keep a streak.

A published report is updated when a material public fact changes (price, close, status, a restated filing). Typos are not a new edition. The original publication date stays. The modified date moves. See corrections on Independence.

← OverviewPublic sources →