Reviews, forums and community answers in AI recommendations

Assistants often draw on reviews, forum threads and community answers because they carry independent opinions from people who have used the product. You cannot honestly write them yourself. You can make a product worth recommending, make it easy for real customers to say so, and take part openly where your buyers ask questions.

This is the part of AI visibility that cannot be shortcut, and for that reason it is hard for competitors to copy.

Why do assistants use reviews and community pages?

Because a question like "what do people actually use for X" is asking for experience, not marketing. Reviews and discussion threads contain that: what worked, what broke, what someone switched from and why. They sit outside any single vendor's control, which makes them a useful source.

They also tend to match how people phrase questions. A forum thread titled with a buyer's problem is closely shaped like the question an assistant is asked.

Which pages are used depends on the question and the assistant. You can see it by testing. The kinds of pages AI assistants cite most lists the usual types and how to find yours.

What do assistants make of mixed opinions?

They tend to summarise. If a discussion contains praise and complaints, the answer may carry both: a recommendation with a caveat, or a recommendation for one kind of user and a warning for another. A single negative thread is unlikely to define you, but a consistent complaint across many sources will show up.

That is useful information. Complaints that recur in reviews and communities are usually real product or service issues, and fixing them improves both your reputation and what assistants read.

What can you do that is honest and effective?

Make the product easy to recommend

This is the unglamorous core. People recommend things that solve a problem well, with fair pricing and good support. No tactic compensates for its absence.

Make it easy for customers to review you

• Ask at a natural point: after a successful outcome, not after a support failure.
• Point to the platforms your buyers actually read. Look at which review sites and communities appear when you test your key questions.
• Keep the process short: a direct link, a sentence of guidance, no pressure on what to say.
• Never offer payment or rewards for positive reviews. Many platforms prohibit it, and it undermines trust.

Respond to reviews

Reply to feedback, including negative, with facts and politeness. A calm, specific reply to a complaint is read by everyone who sees it, and it is evidence you are still active.

Take part in communities openly

If your buyers ask questions in forums or communities, answer them usefully.

1. Read the community's rules before posting.
2. Disclose your connection when you mention your own product.
3. Answer the actual question, including when another option fits better.
4. Contribute outside your own product too. Accounts that only promote are recognised and removed.

Publish things worth citing

Useful research, clear documentation and honest comparisons give communities something to link to. See how to write a comparison page AI assistants trust.

What should you never do?

• Write or buy fake reviews. It is deceptive, it breaks platform rules, and enforcement and exposure are real risks.
• Run sock puppet accounts in forums to recommend yourself.
• Pay for undisclosed endorsements.
• Attack competitors in public threads.
• Try to remove honest criticism you simply dislike.

Beyond the ethics, these tactics are fragile. They depend on not being found out, and a single exposure can become the thing people say about you.

How do you see what is being said about you?

Do two things regularly.

1. Read the sources directly. Search the review platforms and communities where your buyers are, and read what is written about you and your competitors.
2. Check the answers. Ask the assistants about your category and about you. Look at which community pages are cited, and read them. Compare what they say with what you know.

If something factual is wrong, note it and trace the source. How to spot when AI says something wrong about your business gives the routine. If the problem is old information, see how to fix outdated information in AI answers.

How does this fit the rest of the work?

It sits alongside your own site and the pages that list you. A clear homepage, a presence on relevant roundups and real customer voices reinforce each other. See how to get onto the roundup and review pages AI relies on for the listing side.

Progress here is slow and cumulative. Measure it the same way as everything else: a fixed set of questions, a regular schedule and a comparison of like with like. Treat changes as real only when they are larger than ordinary variation.

What if the reviews are mostly negative?

Read them as feedback first. Recurring complaints are usually real, and fixing the cause does more than any reply. Respond calmly, state what changed, and keep doing the work. Assistants summarise what they read, so improvement shows up slowly as new, better-informed opinions accumulate. Do not try to bury criticism. Showing that you listen is part of what makes a recommendation credible.

For a first look at how a site appears to assistants today, try the Citewake checker. Citewake reads third-party pages that answers cite and records what they say about each brand, as set out in the methodology.