What makes an implementation plan credible rather than speculative?

Every proposed intervention should connect an observation to a testable hypothesis and a named delivery owner. Use this template:

We observed [specific repeated result]. We found [verifiable information or access issue]. We propose [change] because [bounded hypothesis]. [Owner] will deliver it by [date]. We will check [deployment evidence] and then assess [outcome] using [unchanged comparison method].

What should a technical SEO and content audit cover?

Require evidence from the actual site and relevant platform documentation. Google’s current guidance states that foundational SEO remains relevant, rejects special AI markup as a requirement and says that llms.txt does not improve visibility in Google Search. It also warns against producing scaled variations primarily to manipulate rankings or generative responses. A supplier proposing these activities should explain the specific non-Google use case or evidence behind its recommendation. (developers.google.com)

For ChatGPT discovery, distinguish search access from training permission. OpenAI documents independent controls for OAI-SearchBot and GPTBot; enabling search discovery does not require a blanket decision to permit training. (developers.openai.com)

Request a prioritised diagnosis covering:

• Relevant crawl, indexing, rendering and inclusion controls.
• Current, accessible information about product capabilities, integrations and limitations.
• Conflicting or outdated product statements.
• Exact third-party pages cited in observed answers.
• Whether the recommended work duplicates an existing SEO, content or PR commitment.

What should a competitor citation analysis prove?

Require the analyst to distinguish four observations:

1. The assistant linked to a page.
2. The accessible page mentioned the competitor.
3. The assistant recommended the competitor.
4. The proposed explanation for how these observations relate.

Only the first three are directly observable in that evidence package. The fourth remains a hypothesis unless supported by a stronger design. Mark inaccessible source pages as unverified rather than inferring what they contain.

For digital PR, require truthful positioning, transparent commercial relationships and editorial independence. FTC guidance on endorsements and AMEC’s ethical guardrails provide reasons to reject manufactured reviews or promotion disguised as independent evidence. (ftc.gov)

How do you check delivery capability?

Ask the people who will do the work to walk through a completed example: diagnosis, approval, deployment, QA and measurement. A hiring signal is not delivery proof. For example, Siege Media’s published Senior SEO/GEO Analyst role combines technical diagnosis, analytics and collaboration with content and development teams; it illustrates the mix of skills worth testing, not proof that any particular engagement has those resources assigned. (jobs.ashbyhq.com)