Building smarter
businesses with
AI agents.

This is for a decision, not a data dump

Most competitor reports answer the easiest question: what do the other companies say about themselves?

Evidence Memo answers the useful question: what should you do next?

Each memo starts with a decision such as:

• Should we enter this category now?
• Can we raise our price without losing the market?
• Which segment can we win?
• Is this competitor's new offer a threat or noise?
• Should we build, partner, acquire, or stop?

If the evidence does not support a confident call, the memo says that. False certainty is worse than an honest stop sign.

How it works: AI does the digging, a human owns the call

This is an AI-native service, and that is the point. The same agent systems I write about every week do the research: sweeping public sources, cross-checking claims, building the evidence tables. I direct the work, verify the sources, and stand behind the recommendation.

You get research depth that used to cost five figures from an analyst firm, delivered in 48 hours, at a price a bootstrapped founder can justify. The methodology is not hidden. It is the pitch.

What arrives within 48 hours

A finished 8-12 page PDF and a clean web version, usually 2,500-4,000 words, with:

1. The call. Proceed, change course, test one assumption, or stop. The recommendation appears first.
2. Decision conditions. The facts that would reverse the call.
3. Market map. The relevant competitors, substitutes, and bundled alternatives. It excludes companies that look similar but do not compete for the same budget.
4. Offer and pricing table. Public prices, packaging, limits, contract signals, trials, and meaningful omissions, time-stamped to the research window.
5. Buyer evidence. Public customer language from product pages, reviews, forums, support documents, job posts, or other attributable sources. Quotes link to the original page.
6. Positioning gaps. Claims everyone makes, claims one company owns, and needs the market is leaving exposed.
7. Recommended move. The segment, promise, price logic, proof burden, and first test.
8. Risk register. The strongest reasons the recommendation could be wrong, ranked by impact.
9. Source appendix. Every source used, with access date. Material claims carry inline links so you do not have to hunt through footnotes.

The evidence standard is simple: a material claim needs a source a buyer can open. Company-controlled pages prove what a company says and sells. They do not prove that customers believe it. Estimates are labeled as estimates. A missing public price is reported as missing, never guessed. Conflicting sources stay visible.

The 48-hour promise

The clock starts when payment and a complete intake are both received. You get the memo within 48 hours, including weekends, unless the checkout page states a temporary closure before you buy.

If a required answer is missing, the clock pauses until it arrives. One round of corrections for factual errors is included for seven days. A change in the decision, market, or competitor set is a new memo.

If delivery is late, you can choose a full refund or keep the completed memo at no charge.

What it costs

• First three completed orders: $295 each
• Orders four and five: $595 each
• Standard price from order six onward: $995

One price covers one decision, one market, and up to five named competitors. Sales tax, if required, is calculated at checkout.

No retainer. No sales call. No hidden research subscription.

What this does not cover

Evidence Memo does not include private database access, paid expert interviews, customer interviews, legal advice, investment advice, technical penetration testing, confidential information, or access obtained by deception. It does not contact competitors while pretending to be a buyer. It does not promise market size where credible public evidence cannot support one.

This is a decision research product. It is not a guarantee that a launch, price change, acquisition, or campaign will succeed.

Who should not buy it

Do not buy if you need a 50-slide market overview, a valuation, a legal opinion, primary customer research, or a memo written to justify a choice you have already made.

Do not buy if your decision depends on private facts you cannot share or evidence that does not exist publicly.

If the intake shows the requested decision cannot be answered responsibly from public evidence, the order is declined and refunded before research begins.

What you get

1. The query panel. Your buying questions, run signed out on all four engines, every answer captured and dated.
2. The verdict per engine. Who each engine names, in what order, in what words.
3. The citation map. The exact domains and pages each engine pulled from.
4. The gap list. Why the engines prefer your competitors: missing pages, invisible pricing, broken schema, absent third-party coverage.
5. The fix list. Every gap paired with the concrete fix and the file or URL to change. Specific enough to hand to whoever edits your site.
6. The re-test plan. How and when to re-run the queries to confirm the answers moved.

The standard

Every answer quoted is a dated capture you can re-run yourself. Engines are quoted, never paraphrased into something more flattering. Where a cause is not provable, the audit says so. If your queries cannot be improved from public evidence, the order is declined and refunded before work begins.

The promise

Delivery within 48 hours of complete intake, weekends included. Late delivery means a full refund or the finished audit free — your choice. One round of factual corrections is included for seven days.

What it does not do: no edits to your site, no guaranteed rankings, no fake reviews, and no magic files. There is no llms.txt switch that makes engines cite you, and anyone selling one is guessing.

$495. One company, up to three competitors, one set of queries.