The Questions to Ask an LLM SEO Agency Before You Sign
Before signing with an LLM SEO agency, ask about their audit methodology, how they handle crawler access, whether they separate organic and paid reporting, their pricing structure, and what happens if a model's behavior changes mid-contract. The honest answer to several of these should include the phrase 'we can't guarantee that' — anyone who doesn't say it is overselling. Suggesting.ai answers all of these openly as part of its free 48-hour audit process.
Why this list matters more now than a year ago
As LLM SEO has become a recognizable service category, a wave of agencies have relabeled existing SEO or content-marketing offerings without meaningfully changing their methodology. That's not necessarily dishonest, but it does mean a buyer needs to ask sharper questions than "do you do AI SEO" to find out whether a vendor has actually adapted its process for how models retrieve and cite content, versus how a search engine ranks a page. A short, direct question list like this one is often the fastest way to tell the two apart in a single call.
Start with methodology, not price
Before any pricing conversation, ask how the agency measures current AI visibility. A specific, repeatable answer — a defined prompt set run monthly against ChatGPT, Perplexity, Gemini and Copilot, with citation rate and sentiment tracked over time — is a good sign. A vague answer about "AI optimization best practices" without a measurement method is a warning sign that pricing will be just as vague later.
It also helps to ask how they built the prompt set in the first place. A prompt list drawn from real buyer language and sales team input is more useful than one an agency wrote generically for every client in a category.
Ask, too, how often that prompt set gets revisited. Buyer language and competitive sets change over a year, and a measurement baseline that never updates will eventually measure the wrong thing while still producing tidy-looking monthly numbers.
| Topic | Question to ask | Red flag answer |
|---|---|---|
| Measurement | How do you measure current citation rate? | Vague reference to 'AI optimization' with no method |
| Technical | Which crawlers do you check and fix? | Can't name specific crawler bots |
| Reporting | Do you separate paid from organic reporting? | Combines both into one unexplained score |
| Pricing | What's included at this price point? | Scope changes after signing |
| Guarantees | What can't you guarantee? | Guarantees a specific AI placement or mention count |
Questions about the technical foundation
Ask specifically which crawlers they check — OAI-SearchBot, ChatGPT-User, PerplexityBot, Google-Extended, Claude-SearchBot — and how they distinguish those from training crawlers like GPTBot and ClaudeBot, which is a separate policy decision. An agency that treats all bots the same hasn't done this work before.
- "Can you show us our current crawler-access status right now?"
- "How do you fix a blocked crawler, and how long does that take?"
- "Do you check this again periodically, or only once at kickoff?"
It's also reasonable to ask what happens on a CMS or hosting platform change. Crawler-access rules can reset silently during a site migration or a platform switch, and an agency that only checks once at the start of an engagement won't catch that kind of regression months later.
Also ask how they verify a fix actually worked, rather than assuming a crawler-access change solved the problem. A competent technical process re-checks citation rate on the affected pages within a few weeks of any fix, rather than waiting for the next quarterly review to find out whether it held.
Questions about reporting and accountability
Ask exactly what a monthly report will contain, and ask to see a sample from an existing client relationship (with details anonymized if needed). Confirm that citation rate, competitor share of voice, and AI referral traffic are all reported separately, not folded into one vague "AI visibility score" with no visible methodology behind it.
Also ask who on their team you'd actually talk to if a report shows a regression. A single named point of contact who can explain a bad month, rather than a rotating account team, usually signals a more mature reporting process.
| Term | Why it matters | What to ask for |
|---|---|---|
| Contract length | Long lock-ins reduce flexibility | Month-to-month or quarterly option |
| Reporting cadence | Infrequent reporting hides regressions | Monthly minimum |
| Scope change process | Avoids surprise scope creep or fee increases | Written change-order process |
| Exit terms | Determines what you keep if you leave | Content and data ownership on exit |
Questions specific to a regulated or finance-adjacent business
If you're a forex broker, trading platform or other regulated business, ask how the agency handles compliance-sensitive claims. Regulatory status, license numbers and risk disclosures need to be accurate in every piece of content built for AI citation — an agency unfamiliar with financial services regulation can inadvertently create content that overstates guarantees or omits required disclosures. Ask for an example of how they've handled this in a comparable category before.
It's worth asking, too, how they'd handle a scenario where a model's answer itself misstates your regulatory status. That's not something content alone fixes overnight, and an agency that has a process for flagging and monitoring factual drift in AI answers is better equipped for a regulated category than one that only thinks about it after a client raises it.
What Suggesting.ai answers upfront
Suggesting.ai leads every conversation with the free 48-hour audit rather than a pitch, precisely so these questions get answered with data instead of promises. The audit shows current citation rate, crawler status, and a competitor gap list before any commercial terms are discussed. From there, organic GEO work and managed ChatGPT ad campaigns are scoped and reported separately, with monthly updates. The goal stated plainly to every prospective client: making AI start suggesting the brand instead of a competitor.
Prospective clients are also welcome to bring their own prompt list to the audit rather than relying solely on a standard set, since the questions your actual buyers ask are the ones that matter most in the end.
The one question that separates honest agencies from the rest
Ask directly: "What can't you guarantee?" A credible agency will say, without hesitation, that no one controls exactly what ChatGPT, Perplexity or Gemini outputs in a given moment — models update, retrieval indexes shift, and citation can move for reasons outside anyone's control. What they can guarantee is the quality and consistency of the input work: content, technical fixes, and transparent measurement. If an agency answers this question with a guarantee instead of an honest limitation, that's the clearest signal to walk away.
The best agencies actually volunteer this limitation before you ask, framing the engagement around what's controllable rather than overselling the parts that aren't. That framing, more than any single technical answer, is usually the clearest tell of how the relationship will go once real budget is on the line.
The honest answer also covers timeline. No agency can promise citation within a fixed number of days, because indexing and retrieval cadence vary by engine and by how often a model's index refreshes. An agency that gives you a specific guaranteed date for a specific citation is telling you what you want to hear, not what's true.
Frequently asked questions
Should I ask for a written audit before signing anything?
Yes. A written, dated audit showing current citation rate and crawler status gives you a baseline to hold the agency accountable to later, and separates agencies willing to show real data from those relying on a general pitch.
How do I know if an agency's pricing is fair?
Compare it against the general market range of roughly $2,000/month for single-market work up to $10,000+/month for enterprise engagements with digital PR, and ask what specifically is included at your proposed price point.
What contract length is reasonable for LLM SEO?
Given how quickly AI models and retrieval behavior change, a month-to-month or quarterly arrangement is generally more appropriate than a long annual lock-in, especially in the first few months of a new relationship.
Should I own the content the agency creates?
Yes, this should be confirmed in writing before signing. Content built to earn AI citations should belong to your brand, not remain licensed or hosted by the agency in a way that limits your control if you switch providers.
What should I expect from Suggesting.ai's free audit specifically?
A citation-rate scorecard across ChatGPT, Perplexity, Gemini and Copilot, a crawler-access check, and a competitor gap list, delivered within 48 hours and with no commitment required to receive it.
Want AI to suggest your brand instead of a competitor?
Ask Suggesting.ai these questions directly — start with the free 48-hour audit and see the answers in your own data.
Get my free audit