Buyer-question map
15 agreed questions across discovery, comparison and buying intent in one target market and one language.
Decision summary
An AI search diagnostic helps you decide what to fix before committing to ongoing delivery. TANTU AI reviews one market, one language and 15 agreed buyer questions across three agreed platforms, with two observations per question: 90 planned observations. The $990 engagement includes an evidence review, prioritized action plan and review call, delivered within 10 business days after complete agreed inputs. Its value is a clearer investment decision: which findings justify a correction, which need more evidence and which can wait. Implementation is scoped separately.
Choose questions that can change a decision
Bring a priority offer, the customers you can serve and a short list of objections from real sales conversations. We use these inputs to choose a balanced set within the 15-question allowance. Each question should reveal something actionable about category fit, evaluation criteria or buying readiness. Brand-only questions can check factual accuracy, but they leave less room to examine discovery by buyers who do not know you yet.
For an illustrative equipment supplier, “Which type of machine suits this material?” and “What site information is needed for a quotation?” address different buying tasks. Five minor rewrites of “best equipment supplier” may add less useful coverage. We agree the wording and allocation before collection, record why each question matters and identify the page or business owner likely to act on a finding. New questions replace or extend scope explicitly; they do not silently increase the sample.
Leave the review with a decision, not just a score
Each priority finding should connect the recorded answer to a source, an observed gap and a practical next action. A demonstrably incorrect specification can justify correcting your page immediately. An unexplained absence in two observations may only justify investigating source quality or repeating research later. These are different confidence levels and should not receive the same recommendation.
The handover separates work your team can implement now, evidence it must supply and opportunities worth evaluating for a managed program. Acceptance means you can trace a recommendation back to an observation and identify an owner and completion check. It does not require the answer to change before the diagnostic is delivered. If the evidence is inconclusive, the useful conclusion may be to resolve the website facts first, defer a new content commitment or scope a narrower follow-up question.