AI is changing how people find you. Give them a reason to choose you.

Start with the finding—and the people it describes

Impact research reports that 47% of its AI-active respondents began their latest product search with an AI response, while 72% still visited a brand or retailer before deciding.

The May 2026 survey covered 1,128 U.S. adults aged 18–75 who used AI and had made at least two recent online purchases. Higher-income households were deliberately oversampled. These are self-reported behaviors in that sample, not a census of shoppers or proof of what happens in your business.

My practical takeaway is to examine the entire buying decision. Someone can hear about you in an assistant, see an independent explanation elsewhere, and arrive at your site already holding a specific question. The page they land on needs to help them answer it.

Write the question that comes before your business name

Start with a task a customer wants finished. A local shop might hear, “How can I get better product photos without hiring a full team?” A consultant might hear, “Who can help us fix slow follow-up without replacing our CRM?” These are illustrative questions, not verified search-volume opportunities.

Collect five questions from conversations you are allowed to use. Remove personal details. For each, identify the decision, the constraint, the evidence you have and the next useful action. If every answer becomes “book a call,” you may be asking for commitment before explaining enough.

Build one page that helps a person decide

A useful offer page tells people who it serves, the work included, the inputs needed, the boundaries and what happens next. Explain cases where the offer is unsuitable. Show a real example with permission, or label a demonstration as fictional.

For a product-photo service, that might mean listing the number of products, required source images, review rounds and delivery process. Do not imply that generated lifestyle scenes prove how a real product performs. Give buyers a sample they can evaluate before they enquire.

Organize the page around questions rather than a list of AI tools. Your customer is trying to make a decision about a result.

Let other people contribute independent evidence

The Impact partnership resources are useful prompts for thinking about publishers and creators. Our recommendation is to find a few relevant people who can actually examine your work. Offer clear information and permission to be candid. A partner repeating your sales copy provides little new evidence.

Owned sites can explain different parts of a topic, but links among your own properties are not independent endorsements. Give each page a distinct job and disclose commercial relationships where they affect a recommendation.

Run a small visibility and understanding check

1. Write down five buyer questions and the date.
2. Inspect answers on two relevant AI services, keeping the wording and settings consistent.
3. Record the answer, cited URL, factual errors and whether a reader could take the next step.
4. Ask three people from the intended audience to use the page and explain what they think they would receive.
5. Fix the most consequential misunderstanding and repeat the check.

These sample sizes are manageable learning exercises, not statistical guarantees. A mention in an AI answer is not a lead, and a lead is not a sale. Track enquiries and completed work separately.

Make the next decision smaller

If people understand the offer but do not need it, revise the offer or audience. If they want it but cannot find proof or understand the process, repair that gap. If the next step fails, fix the handoff before expanding distribution. AI discovery can be part of your plan without becoming an excuse to postpone these fundamentals.

Put it to work

Choose one offer and one buyer question today. Write the missing answer, add evidence you can stand behind, and ask someone in your audience to test the page. Use the buyer-intent guide to connect the question to a useful offer.

Sources and editorial context

Reviewed November 3, 2026. This is independent editorial analysis. The exercises, sample briefs and decision rules are our recommendations; they are not results reported by the sources or measured client outcomes.

• Impact research
• Survey methods
• Partnership resources