AI SEO Services: What to Get in Writing Before You Pay

The second job matters because clicks were scarcer on Google pages with an AI summary. In Pew Research Center's browsing data from 900 U.S. adults in March 2025, 18% of Google searches produced an AI summary. Users who saw one clicked a traditional result link in 8% of visits, against 15% for users who didn't. Pew reports an observed difference and doesn't claim the summaries caused it, but it's reason enough to want a report on what the summary says about you as well as where you rank.

Key points

• "AI SEO" on a proposal can mean either job, so ask which one each line item is and what report will show it worked.
• For Google AI Overviews and AI Mode, Google says no special optimization is needed beyond being indexed and eligible to show with a snippet. That part is ordinary SEO.
• For ChatGPT, Gemini, Perplexity and Claude, ask for a dated report, engine by engine, that splits questions naming you from questions that only describe your service.

Which AI SEO service are you actually being sold?

	AI-assisted SEO	Work on AI answers (often sold as AEO or GEO)
What the AI is for	Producing SEO work faster	The place your business needs to appear
Typical line items	Keyword research, content drafts, technical audits, reporting	Buyer questions run on named engines, a record of who gets named and cited, recommended fixes
How results are checked	Google rankings and clicks	Whether each engine names or cites you, by question type
Usual failure	Many pages published with nothing added for readers	One total that hides the questions where you're missing
Google's rule that applies	Spam policy on scaled content abuse	For AI Overviews, normal Search eligibility and nothing extra

Use one rule to decide which comes first. If your pages aren't indexed in Google or can't show with a snippet, start with ordinary SEO, because Google's guide to AI features requires both before a page can appear as a supporting link in AI Overviews or AI Mode. If they already show in Google but you don't know what ChatGPT or Gemini says when a buyer describes your service without your name, measure that first.

If AEO and GEO are new terms to you, AEO vs SEO sorts out which means what.

Hypothetical example: a two-office dental practice pays for "AI SEO." Each month's report shows 40 new AI-drafted blog posts and higher Google rankings for 12 keywords. Nothing in it says whether ChatGPT names the practice when someone asks for a dentist nearby. The practice bought the first job and assumed it was getting the second.

What should AI-assisted SEO deliver?

Pages that help a reader, checked by a person, at a pace you can review. Google's guidance on using generative AI content says, "Generative AI can be particularly useful when researching a topic, and to add structure to original content." The same page warns that using AI tools "to generate many pages without adding value for users" may violate Google's spam policy on scaled content abuse.

So ask any provider selling this job:

• Who reads each page before it goes live, and what do they check it against?
• What does each page add that the pages already ranking for that search lack?
• How many pages a month are planned, and why that number?

What should a report on AI answers show you?

Ask for these five things in writing before you pay:

1. The questions. The buyer questions to be tested, split into ones that name your business and ones that only describe a need, such as your service plus your town.
2. The engines and conditions. Which engines and models, the dates, how many times each question was asked, and whether answers came through an API or the consumer app. API answers collected under recorded conditions are not the consumer apps, so the report has to say which it measured.
3. Named and cited, counted apart. Being named in an answer and having your site linked as a source are different results.
4. An access check. OpenAI's crawler documentation says, "Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers." Someone should open your robots.txt and confirm it doesn't block that crawler by accident.
5. Fixes tied to findings, and a recheck. Each recommended change should point to a question where you were missing, and the same questions should be asked again later on the same engines.

How can a large citation count still mislead you?

By adding together questions that name you and questions that don't. In a practice audit Vector Index delivered at no charge in September 2026, run on ChatGPT, Gemini and Perplexity through their APIs, the business's main site was cited 411 times. Of those, 239 came from questions that named the business. On questions that described a buyer's need without naming anyone, the same site was cited 8 times across 292 answers. Vector Index reports those two slices on its services page, linked at the end of this article, and doesn't break out the rest.

A total of 411 reads well in a monthly report. But a buyer who doesn't know your name yet asks the second kind of question, and there this site drew 8 citations in 292 sampled answers (citations counted in answers, not buyers who saw the site). That audit's setup differs from the one Vector Index sells now, so treat it as one dated reading, not a benchmark. Either way, ask for the two kinds of question reported apart, and read the second number first.

Why does tracking one engine miss so much?

Because in Vector Index's samples, the engines cited different sites. When Vector Index put 16 buying questions about Cincinnati businesses to ChatGPT, Gemini and Perplexity through their APIs on 24 and 25 August 2026, the three engines' answers cited 63, 98 and 181 different domains. Of the 246 domains cited in total, 18 appeared in answers from all three.

The local citations study shows the difference at the level of single sites. Across 168 answers to questions that named Cincinnati or said "near," sampled on 24 and 25 August and 16 September 2026, ChatGPT cited bbb.org in 14 of its 68 answers. Gemini cited bbb.org in none and reddit.com in 14. The study calls these readings on the days sampled, not rates, taken through each engine's API with web search on, not the consumer apps. So if a provider tracks one engine, the sources it tells you to fix may not be the ones another engine cites. Ask for every finding broken out by engine.

What can't any AI SEO services provider promise?

A position. Google's advice on hiring an SEO says, "No one can guarantee a #1 ranking on Google," and, "If they guarantee you that their changes will give you first place in search results, find someone else." Nobody controls what ChatGPT or Gemini writes either, and the local study notes, "Engines and models change; a later sample may differ."

A provider can commit to a defined method, a dated baseline, fixes tied to it and a date to measure again. A proposal vague on those and specific about outcomes is a warning.

Where should you start with AI SEO services?

If you're comparing AI SEO services, start with what the engines say today when a buyer describes your service without your name, because a ranking report doesn't measure that. Open Vector Index's answer engine optimization services page and read the audit scope. The audit asks 25 buyer questions, each four times on five named engines, under recorded conditions. It ends with a 30-day roadmap and the wording for each change, which you or your existing web provider can act on, and after 90 days five of the original questions are checked again so you can see what changed. You get a measured baseline before you pay anyone to make changes.