How To Brief An AI SEO Agency Properly

One test separates a report written to inform from one written to reassure. Read it and try to write down a question it does not answer. In a good report you will find several, because it contains enough specifics to make new questions obvious. In a padded one you will struggle, not because everything is covered but because there is nothing specific enough to interrogate.

Refusals matter. A report that only contains successes is either describing a suspiciously easy month or omitting the parts that did not work, and the omitted parts are usually where the useful information is.

Ask for one change and see what happens: request the raw answers and the run counts. An agency doing the work sends them the same day, since they already exist. One that does not will explain why the format makes that difficult. llm seo

A Claim About Causation, Stated as a Claim The most valuable paragraph in the report is the one that says what moved and why the agency believes their work caused it, phrased as a judgement rather than a fact.

Include the constraints too. The job size you turn down, the sector you do not serve, the situation where a competitor is genuinely the better answer. Those are the statements that get quoted, and an agency will not invent them for you.

One section most briefs omit is worth adding: what has already been tried and what happened. Agencies frequently propose work that was done two years ago and abandoned, because nobody told them. Listing previous efforts, including the ones that failed, saves a month and signals that you will be a straightforward client to work with.

The reasonable reading is that ranking gets a page considered while quotability and corroboration decide whether it is used. Treating a strong search position as an entitlement to appear in answers is the mistake that catches out established brands most often.

The honest framing first: nobody outside these organisations knows the selection logic, and the systems change without announcement. What follows is drawn from observable behaviour, visible citations and published research, which supports useful generalisations and does not support precision.

Good versions read like this: mention rate on evaluation prompts rose from two in fifteen to six in fifteen, which we attribute to the three directory corrections completed in week two, though a competitor also stopped publishing during the same period.

Say What You Will and Will Not Approve Publishing constraints kill more engagements than capability gaps. If every page needs legal review with a three week turnaround, say so, because it changes what is realistic and what should be prioritised.

Two implications follow regardless of which system you are studying. Being findable by the underlying search step is necessary, and being worth quoting once fetched is what decides whether you are used. Almost everything actionable sits in those two requirements.

The better approach is to keep them, correct the facts, date them honestly, and make clear how they relate to the present. A page that says plainly what it documents and when is more useful than one quietly rewritten to look current.

We also know the picture is unstable. Retrieval strategies are revised without announcement, and a method that explained answers well six months ago may explain them poorly today. Anyone selling certainty here is selling something they do not have.

Work Completed, in Countable Units Listings claimed, with names. Errors corrected, with the source and what was wrong. Pages published or rewritten, with URLs. Technical changes made, with dates. Outreach attempted and its outcome, including refusals.

This section sounds procedural and it is the foundation of everything after it. A prompt set quietly edited between runs makes every trend line in the document meaningless, and it is the easiest way to manufacture improvement without doing anything.

What Ranking Does and Does Not Buy You Ranking still helps, because the retrieval step usually starts with a search. But it buys far less than people assume. Ahrefs examined 15,000 long-tail prompts across four assistants in July 2025 and found roughly 80 percent of cited pages did not rank for the original query at all, with about 12 percent in the top ten.

The lesson generalises to any brand whose name is short, generic or ambiguous. The correction is not clever, it is repetitive: pick one written form, use it everywhere, and pair it with a descriptive phrase so that a mention alone is never the only clue about what it refers to.

Keep the brief to something you would be willing to send to three suppliers unchanged. The temptation is to tailor each one, which feels attentive and makes the resulting proposals impossible to compare. Identical briefs produce differences that reflect the agencies rather than the instructions, which is the entire point of asking more than one.

This variability is the main practical trap. Testing without web access and concluding you are invisible measures the training corpus rather than current retrieval, and the two can disagree sharply. Record which mode you used with every run.