What Should an Agency Do When a Client’s Cited Pages Are Technically Correct but Commercially Weak?

Why do cited pages fail commercially?

• They answer early-stage questions. "What is a heat pump?" attracts people who are learning, not buying.
• The AI answer is enough. For informational searches, most people get what they need without clicking. Seer Interactive found organic click-through on informational queries with AI Overviews averaged 0.70% when the brand was cited and 0.52% when it was not, in the third quarter of 2025.
• No next step. The page ends without linking the topic to the client's service, area or offer.
• No reason to choose the client. No proof, prices, credentials or examples.
• The wrong audience. DIY readers or students find the page, not buyers.
• Citations without names. A Semrush and Growth Memo study found sites were often cited without being named in the answer, so readers do not connect the information to the client.

How do you diagnose a weak cited page?

1. List the prompts that cite each page, and classify them by buying stage: learning, comparing or ready to buy.
2. Check AI referral traffic to the page in analytics, and what those visitors do next: bounce, read more or enquire.
3. Look for the next step on the page. Is there a link or offer that fits a reader who just learned the answer?
4. Compare with pages that do convert, usually service, pricing, comparison and case study pages.
5. Check whether decision-stage prompts cite the client at all. Often they do not, which is the bigger problem.

What should I fix first when AI cites informational pages but visitors never request a quote?

Start with the pages that are cited most, and give them a bridge to the business:

• Add a short, relevant next step after the main answer, such as "If your system is over ten years old, a heat pump survey will tell you whether it is worth replacing. We cover [area]; surveys are free." The step should follow from the question.
• Link to the decision pages that answer what comes next: prices, service areas, how the process works, and case studies.
• Add one piece of proof on the page, such as a short example or a credential, so readers see the client as someone who does this.
• Qualify the reader: "This guide is for homeowners in [region]. For commercial buildings, see [page]."

Then build or improve the decision-stage pages that buyers' later questions need. Those pages turn into enquiries. How decision-stage pages drive leads from AI answers is covered in whether decision-stage pages drive inbound leads from AI answers.

How do I explain why a cited page did not generate a lead?

Be direct and give evidence:

• A citation is not a visit. Most people read the AI answer and move on. For informational questions, that is normal.
• The question was early-stage. People asking "what is" questions are often still learning, and many are weeks or months from buying, if they buy at all.
• The page was not built to convert. Say what you will change, such as adding a next step or linking to decision pages.
• Some influence is invisible. Readers may remember the client's name and search for it later, or arrive as direct traffic. Branded search trends and "how did you hear about us" answers help show this.
• Conversion rates vary by source. Published figures conflict: a 2025 working paper on nearly 1,000 ecommerce sites found ChatGPT referrals converted worse than Google organic search, email and affiliate traffic, but better than paid social, while other studies have reported AI visitors converting better on the sites they measured. The picture differs by sector, which is why each client's own numbers matter most.

Then show the plan: which buying questions you are targeting, which pages will serve them, and what you will measure.

Which pages should the agency build next?

Look at the questions buyers ask once they have learned the basics, and make sure the client has a page that answers each one:

• Cost pages: typical prices or ranges, what changes the price, and financing or grants.
• "Is it worth it?" pages: honest comparisons of options, including when the client's service is not the right choice.
• Process pages: what happens from first contact to completion, how long it takes and what the customer needs to do.
• Service area and location pages, for businesses that serve specific places.
• Case studies with specifics: the problem, the work, the result and the cost range.
• Comparison pages for the alternatives buyers weigh, such as a competitor, a different approach or doing it themselves.

These are the pages that answer the questions closest to an enquiry. Link each informational guide to the one or two that follow naturally from it.

How should agencies report this?

Separate the stages, so citations are not mistaken for revenue:

Agree these measures with the client at the start. It avoids the conversation where a client sees many citations and no leads, and concludes the work failed.

What might an illustrative case look like?

Consider an illustrative agency whose home-energy client was cited in dozens of AI answers about how heat pumps work, but received no quote requests from AI referrals. The cited guide ended with a newsletter sign-up. Decision questions such as "heat pump installer cost in [county]" cited competitors and a price comparison site. The agency added a short survey offer and links to the client's installation, pricing and grants pages at the end of each cited guide. It also built a pricing page with typical costs by home size. It reported citations by stage alongside enquiries. Three months later, next-step clicks from the guides had tripled, and the pricing page was cited for cost questions and produced the first AI-referred quote requests.

What do people ask about commercially weak citations?

Should we stop creating informational content?

No. It builds familiarity and supports later buying questions. Just make sure it connects to the business and is not the only content.

How many leads should AI citations produce?

It depends on the sector and the prompts. Set expectations from the client's own data after a baseline period.

Can a next step look too salesy on a guide?

Keep it relevant and brief. A step that follows naturally from the question helps readers.