How to Choose an AI-First SEO Agency: 9 Questions to Ask

Almost every SEO agency in 2026 claims to be “AI-first” or “AI-powered”. Most of them are not. Some have just bolted ChatGPT onto a 2018 process and renamed the deck.

This post is the nine-question checklist to use on a discovery call. Each question has a “good answer” and a “red flag answer” so you can tell, in 30 minutes, whether the agency is actually AI-first or just using the label as marketing.

Why Vetting Matters More Than Ever

Hiring the wrong agency now costs more than it used to. AI search is reshaping commercial visibility in months, not years. Every quarter you spend with an agency that ignores AI engines is a quarter your competitors who do not are pulling ahead.

The good news: spotting the difference is fast if you ask the right questions.

Question 1: How Do You Use AI in Audits?

What you want to hear: “We use code-aware AI agents to parse log files, run gap analysis across all URLs, generate prioritised technical recommendations from raw GSC data, and validate schema at scale. We have built our own audit pipeline that connects directly to Ahrefs, GA4, Search Console and Google PageSpeed, scrapes the site for structure and AI-readiness signals, and queries Perplexity, ChatGPT and Gemini in the same run — so the audit is cross-referenced across SEO, technical and AI-citation data in a single brief. A senior strategist supervises and interprets, but the AI handles the volume work.”

Red flag: “We use ChatGPT to summarise things sometimes.” That is using AI like a calculator, not building it into the process.

Question 2: How Do You Measure AI Search Visibility?

What you want to hear: A specific tool stack with honest scope. “We use a combination of tools and manual validation. We track citations and brand mentions across AI engines using Ahrefs Brand Radar. We monitor Google AI Overviews appearance for target queries. We benchmark answer share against competitors quarterly and are piloting deeper attribution work as the engines expose more data. Here is a sample report.”

Red flag: “AI search is too new to measure reliably.” The tools exist and the methodology is evolving quickly. Serious agencies are already measuring directional signals — citations, answer share, brand mentions in AI engines. If an agency is not tracking any of that, they are almost certainly not optimising for it either.

Question 3: What is Your Entity SEO Process?

What you want to hear: “We map your brand, products, services, team and customers as entities. We optimise structured data (Organization, Person, Service schema), strengthen Knowledge Graph signals, and build entity authority through citations in industry databases. The goal is making AI engines confidently cite you when answering questions in your category.”

Red flag: Vague mention of “schema markup” without explaining how entities differ from keywords or how they tie to AI search visibility.

Question 4: How Long Until First Signals?

What you want to hear: Realistic, specific timelines. “Technical wins in 30-60 days. Content and authority gains by 90 days. Solid AI visibility shifts in 4-6 months.”

Red flag: Either “Page one in 30 days, guaranteed” (impossible and dishonest) or “12-18 months minimum” (a sign their process is slow, not that SEO is slow).

Question 5: Who Owns My Account Day to Day?

What you want to hear: A named senior strategist with their experience and previous clients. “Maria is your dedicated strategist. She has 8 years of SEO experience including law firms and SaaS. You will speak to her every two weeks at minimum.”

Red flag: “We have a team approach” with no named senior. That usually means an account manager fronts the relationship while juniors do the work.

Question 6: What Does Your Reporting Include?

What you want to hear: Rankings, traffic and conversions PLUS AI citation share, brand mentions across AI engines, answer share for target queries, and entity strength signals. Ask to see a real (anonymised) sample report before signing.

Red flag: A 2022-style report with rankings, traffic and a screenshot of GSC. If they cannot show you AI metrics, it is fair to assume they are not tracking them.

Question 7: How Do You Price Your Work?

What you want to hear: A clear answer on where the AI-first economics actually show up. That can take three forms: lower cost per output (the same audit delivered in 6 hours of senior time instead of 30 hours of mixed senior+junior time — fewer hours of more expensive talent producing the same deliverable), faster turnaround (technical fixes shipped in days, not sprints), or broader scope at the same price (entity SEO, AI search measurement and content production bundled where competitors charge separately for each). At least one of those should be visible in the proposal.

Red flag: Same price, same deliverables as a 2022 retainer, with “AI-first” added to the deck cover. The label is marketing.

Question 8: Show Me a Real Result

What you want to hear: Specific case studies with numbers. “We grew this client to 4,800% organic traffic in 18 months.” “We took this brand from 0 to 47% answer share in their category.” “We rebuilt this firm”s technical foundation and Core Web Vitals went green within 4 weeks.”

Red flag: Vague claims with no numbers. “We helped many clients grow their visibility.” Cool, but any agency can say that.

Question 9: How Do You Produce and Validate Content in an AI-First Workflow?

What you want to hear: A clear human + AI split. “AI handles research, outlining, first drafts, schema generation and surface QA. A senior writer or strategist with subject expertise then rewrites for voice, accuracy and originality. We add proprietary insights, examples and opinions that AI cannot generate from training data alone. We run plagiarism and AI-detection checks, but the more important filter is whether the piece would be useful to a competitor as written. If yes, we have not done our job.”

Red flag: “Our AI generates publishable content end-to-end” or “we have proprietary AI that writes better than humans.” Both usually translate to thin, indexable-but-uncitable content that hurts long-term authority. AI-first does not mean human-out.

Red Flags to Watch For

Beyond specific answers to those nine questions, watch for these broader signals that the agency is not what they claim:

• Their own SEO is weak. If their site does not rank or get cited, why would yours?
• No named team on their site. Mature agencies show their senior people. Faceless agencies are usually account-management shells over offshore delivery.
• Pushing a 12-month contract on day one. Confidence in your work means confident in month-to-month terms.
• Heavy “AI” branding with no demos. If they cannot show you AI tooling in action, they probably do not have it.
• Promising specific rankings. “Page one for ”lawyer london” guaranteed.” Anyone promising specifics is either lying or about to use risky tactics.

Score Your Discovery Call

Run through the nine questions and the broader red-flag list above, then tally how many red flags came up in the agency’s answers. A simple rule of thumb:

• 0–2 red flags — Strong agency. Worth a second meeting and a paid pilot.
• 3–5 red flags — Proceed with caution. Push back on specifics, ask for a sample report and a named senior strategist before signing anything longer than a 90-day pilot.
• 6 or more — Walk away. The “AI-first” label is doing the heavy lifting, and you will end up paying retainer prices for a 2022 process.

For more context on what an AI-first agency actually does behind the scenes, see our deep dive on what an AI-first SEO agency means.

If you want a real example of how this looks in practice, book a free discovery call. We will walk through the questions above with our actual answers, show you a real reporting sample, and you can decide for yourself whether what we describe matches what we deliver.