Short answer

Commercial intent means a buyer is evaluating options. Transactional intent means they want to take the next step. In AI search, the useful distinction is the decision still unresolved: which option fits, or how to start. Build your page around that missing answer.

Read the decision behind the prompt

A prompt can contain both comparison and action. “Find a CRM for a five-person team and book a demo” asks for a shortlist and a transaction. Classifying it as one keyword category loses the handoff between those jobs.

Use the categories as editorial tools, rather than claims about what an AI system secretly thinks. Record the buyer’s constraints, the uncertainty they want resolved and the next step they are willing to take. If the prompt does not reveal readiness, label it uncertain.

Match the question to the right source page

An illustrative page map for a B2B software company
Buyer questionIntentUseful page
What does a CRM do?UnderstandA plain-English explanation with examples
Which CRM fits a five-person team?CompareCriteria, alternatives and explicit limits
Does it sync with our accounting tool?VerifyIntegration details and support boundaries
Can we try it this week?ActTrial terms, setup requirements and a working action

The example describes content responsibilities, not measured search demand. A single page can resolve adjacent questions when they share the same decision. Split pages when the reader needs a different answer or action, not just a synonym.

Give comparison pages a fair decision rule

A comparison page helps when it explains who each option fits and why. State your criteria before naming a preferred option. Include material exclusions, switching effort, integrations and the date of any checked pricing.

“Best” without a use case leaves the buyer to invent the criteria. “A fit for teams that need an existing accounting integration and a guided rollout” is more useful, provided the product actually supports those requirements. Link claims to product documentation or inspectable evidence.

Make action possible without hiding conditions

A transactional page should state what happens after the click. A demo request is different from a confirmed booking; a quote request is different from a published price. Show the relevant conditions near the action.

For a service business, that might mean service area, eligibility, starting scope and response expectations. For software, it might mean trial duration, setup steps and who can cancel. Check the form on a phone and verify the confirmation yourself.

Treat visibility and buying intent as separate observations

Being cited in an answer does not establish that the reader wants to buy. Record mentions, citations, visits and qualified inquiries separately. Compare pages serving similar decisions before deciding that one intent performs better.

Google’s guidance for AI search features retains ordinary Search eligibility requirements. An intent label is an editorial aid, not a special markup requirement or a promise of inclusion.

Sources and editorial scope

Published October 20, 2026. This is Arrow AI’s editorial method. Examples and scores are illustrative, not measured demand, customer results or AI ranking factors. The linked documentation provides platform context.

The Arrow read

Connect the answer to the next decision

Arrow AI connects public AI visibility with the source pages buyers use to evaluate a business. Start with the free GEO audit, then use this intent map to choose whether the next improvement belongs on a comparison, integration, pricing or booking page. Run the free GEO audit.