The five most important questions to ask an SEO agency before hiring are: (1) "Walk me through how you would prioritize the first 90 days for our business specifically." (2) "Can you show me a real client report from the last quarter, including both Google rankings and AI citation share?" (3) "How do you measure success, beyond keyword rankings and traffic?" (4) "Which AI surfaces (ChatGPT, Perplexity, Gemini, AI Overviews, Claude) do you optimize for, and how is that work different from classic SEO?" (5) "Who specifically will work on our account, what are their seniority levels, and how many other accounts do they handle?" Their answers reveal more in 20 minutes than any pitch deck ever will.

The full checklist below covers 15 strategy and process questions, eight AI-search-specific questions, a red flags section, and a proposal-comparison framework. Use it on every shortlist call. Buyers who run this checklist consistently end up with stronger agency partnerships and avoid the most expensive hire of all, the wrong one.

Why vetting an SEO agency matters more in 2026

The SEO buying decision changed in 2025, and most vetting checklists have not caught up. Two structural shifts make agency selection harder than it was three years ago.

The first is the AI search overlay. Google AI Overviews now appear on approximately 48% of tracked queries, per BrightEdge data from February 2026. ChatGPT, Perplexity, Gemini, and Claude each take a meaningful share of branded and category-defining searches that used to go straight to Google. An agency that runs a 2020 playbook (rankings, links, on-page) without an AI search layer is optimizing for half the surface area buyers actually search across.

The second is the validation problem. SEO has always had a quality gap between top-tier and commodity providers, but the gap has widened because AI-search work is genuinely new and most agencies are still learning it in public, often on client budget. The legacy "what tools do you use, what's your link-building strategy" checklist no longer separates competent agencies from underqualified ones. Asking the right questions about both classic SEO and AI visibility is the only way to tell which agencies have built the muscle and which are using AI as a marketing label.

Hiring the wrong agency now compounds faster than it used to. Cited brands inside AI Overviews earn approximately 120% more organic clicks per impression than uncited brands on the same queries, per Seer Interactive's March 2026 update. That citation premium is not optional. Each quarter spent with an agency that ignores AI surfaces is a quarter your competitors who do not are pulling ahead, and the gap is hard to close once it opens.

Before any agency call, get clear on three things internally: what business outcome you want from SEO (qualified leads, ecommerce revenue, brand discovery, all three), what your current performance baseline looks like across Google and AI surfaces, and what budget range you can actually sustain for 12 months. Walking into a vetting call without those three pieces guarantees you will be sold on whatever the agency does best, not whatever your business needs.

15 questions to ask about strategy and process

The 15 questions below cover strategy depth, execution model, measurement, and partnership fit. Run them on every shortlist call in the same order so you can compare answers across agencies cleanly.

1. Walk me through how you would prioritize the first 90 days for our business specifically.

You are listening for genuine specificity tied to your business, not a generic roadmap. A strong answer references your actual category, your visible competitors, and a sequence that starts with a baseline audit, then technical fixes, then content velocity. A weak answer is a templated four-phase methodology that could apply to any client.

2. Who specifically will work on our account, what are their seniority levels, and how many other accounts do they handle?

The single most common buyer regret in SEO agency relationships is signing for senior talent and getting served by junior account managers. Ask for named team members, their job titles, their years of experience, and the number of accounts each manages. Eight to 12 accounts per strategist is healthy. Above 20 is a red flag.

3. Can you show me a real client report from the last quarter, including both Google rankings and AI citation share?

If the agency cannot show a real report (anonymized is fine), it does not have the operational discipline to produce one for you. Look for: ranking velocity, organic traffic trend, conversion or revenue attribution, AI citation share across at least three AI platforms, and clear quarter-over-quarter deltas. Generic vanity charts (impressions up, no context) are a red flag.

4. How do you measure success, beyond keyword rankings and traffic?

A strong answer ties measurement to business outcomes: qualified pipeline, revenue attributed to organic, AI citation share in target query sets, share of voice in branded searches, and customer acquisition cost trend. A weak answer stops at rankings and traffic, which are leading indicators, not outcomes.

5. How do you decide what to write and ship next?

You are testing for content prioritization discipline. A strong answer references SERP gap analysis, AI citation gap analysis, internal content velocity, and revenue potential by query. A weak answer is "we follow our content calendar." Calendars are not strategy.

6. What does your link-building approach look like, and how do you avoid penalties?

Quality matters far more than volume in 2026. Acceptable answers include digital PR, editorial outreach, broken-link campaigns, expert citations, and HARO-style sourcing. Red-flag answers include paid placements at scale, link exchanges, or any reference to private blog networks. Ask whether the agency has had a client penalized in the last three years and how they handled the recovery.

7. How do you handle technical SEO, and which engineers or developers do you work with?

Technical SEO is the floor of the discipline. The agency should be able to describe how they audit crawlability, indexation, site speed, schema, internal linking, and JavaScript rendering. Bonus signal: they ask about your dev cycle, ticketing process, and which CMS you use before answering.

8. How do you stay current with Google algorithm changes?

You are testing the agency's epistemic discipline, not their pattern memorization. A strong answer mentions Search Central documentation, named analysts they follow, conference attendance, internal post-mortems after major updates, and how the agency translates changes into client action. A weak answer is "we monitor industry news."

9. What does your reporting cadence and dashboard access look like?

You should expect a live dashboard, monthly written reports, and at least one strategy call per month. Less than that signals an under-resourced engagement. More than that often signals an agency that bills for meetings rather than work.

10. How do you handle communication, account management, and approvals?

Ask for the exact response SLA on email and Slack, the meeting cadence, and the approval workflow for content, technical changes, and link campaigns. Vague answers ("we'll figure it out together") are a red flag at the contract stage.

11. What is your typical client tenure and churn rate?

Healthy SEO agencies retain mid-market clients for 18 to 36 months on average. Less than 12 months suggests delivery problems or misaligned expectations. The agency should answer this honestly and without spin. If they dodge the question, ask why.

12. What kind of client is a bad fit for you?

The right answer is specific. Wrong-fit clients usually include very low budgets, immediate-results expectations, no internal content or dev support, and businesses in restricted verticals (gambling, adult, certain financial services). An agency that says "we work with everyone" either does not know its own positioning or is desperate for revenue.

13. What is your typical content production capacity per month, and how do you manage quality?

Ask for the actual number of pieces per month at your retainer level, the editorial process (briefs, drafts, edits, fact-checking), and how AI is or is not used in production. Reasonable mid-market output is 4 to 12 pieces per month depending on depth. Honest agencies disclose their AI usage policy; evasive ones often hide it.

14. How do you handle migrations, redesigns, or platform changes?

Most mid-market brands will face at least one site migration during a three-year agency engagement. Ask whether the agency has run migrations on your CMS, what their migration checklist looks like, and how they protect rankings during the cutover. If they have never migrated a site like yours, treat that as missing experience.

15. What are the contract terms, and can we exit if performance does not meet expectations?

You want month-to-month or three-month terms, written cancellation language, and a clear definition of what gets handed over (analytics access, content drafts, link inventories, custom dashboards) on exit. Twelve-month lock-ins with no performance clauses favor the agency, not the client.

Questions specific to AI Overviews, GEO and AEO

The eight questions below test for genuine AI search capability versus marketing-label use. Buyers who skip these in 2026 are paying for half the discipline.

16. Which AI surfaces do you optimize for, and how is that work different from classic SEO?

A strong answer names ChatGPT, Perplexity, Google AI Overviews, Gemini, and Claude, and explains that each surface uses different retrieval signals: ChatGPT skews to Wikipedia and Reddit, Perplexity is real-time and citation-first, AI Overviews run on Google's index, Gemini overlaps heavily with Google Search, and Claude has its own quirks. A weak answer is "we cover AI search" with no platform-level distinction.

17. Can you show me a citation report for a current client across the major AI platforms?

The agency should produce an anonymized report showing the client's citation share across at least three AI surfaces, with quarter-over-quarter deltas and competitor benchmarks. If they cannot show this, they are not measuring it, which means they cannot improve it.

18. How do you track citation share over time, given that AI answers are probabilistic?

You are listening for a real methodology. A strong answer mentions running the same target queries on a recurring schedule (weekly or biweekly), tracking appearance frequency rather than fixed rankings, and using either a dedicated platform like Passionfruit Labs, Profound, or Otterly, or a structured manual testing process. Weak answers treat AI visibility as a single static metric.

19. How do you structure content for retrieval and citation by LLMs?

A strong answer references direct-answer paragraphs in the first 200 words of every piece, FAQ schema where appropriate (without overdoing it), entity-consistent naming across the site and the broader web, and clear semantic HTML. A weak answer focuses on "AI-friendly format" without specifics or pushes llms.txt as a primary lever (Google has explicitly said llms.txt does nothing for its AI features, per its May 2026 documentation).

20. What is your view on the relationship between SEO, AEO, and GEO?

You are testing for nuanced framing. A strong answer acknowledges that Google has officially called GEO and AEO "still SEO" for its own AI features while also explaining that ChatGPT, Perplexity, and Claude operate on different retrieval signals that require additional work. The breakdown is covered in our explainer on whether GEO is still SEO. A weak answer either dismisses GEO entirely or insists it is a separate magical discipline.

21. How do you handle third-party platforms (Reddit, YouTube, Wikipedia, G2) that AI surfaces cite heavily?

A strong answer includes a digital PR strategy, a Reddit or community engagement plan, a Wikipedia presence audit, and a review-site optimization plan for G2, Capterra, and TrustRadius if you are B2B. A weak answer focuses only on owned content. AI surfaces cite owned content at a fraction of the rate they cite third-party platforms, which means an agency that ignores the latter is leaving most citations on the table.

22. What is your stance on AI-generated content, and how do you ensure quality?

A strong answer is specific about where AI is used (research, outlining, first drafts of certain content types) and where it is not (final voice, expert insight, fact-checking, anything passing as first-hand experience). The agency should reference Google's E-E-A-T guidance and have an internal editorial review process. A weak answer is either total denial of AI use or full delegation to AI without human review.

23. How do you measure ROI from AI search visibility?

A strong answer separates citation share (the leading indicator), referral traffic from AI surfaces (often under 5% of total traffic but with a 4x to 5x conversion premium per Washington Post data reported by Digiday), and revenue attribution where it can be measured. A weak answer claims direct attribution that is not technically possible in the current measurement environment.

Red flags in agency answers

The patterns below repeat across underperforming agencies. If you hear three or more in a single discovery call, walk away.

The agency guarantees specific rankings or AI citations. No reputable provider does this because none of the algorithms or retrieval systems support deterministic guarantees. Guarantee language is either dishonest or naïve.

The agency cannot name the people who will work on your account, or named contributors change between the pitch and the kickoff call.

Pricing is implausibly low for the stated scope. A $500 per month retainer cannot fund a senior strategist, content production, technical work, link building, and AI search tracking simultaneously. The math does not work.

The agency presents case studies without verifiable client names, before-and-after dates, or specific traffic and revenue numbers. Generic "200% increase in traffic" claims with no baseline are decoration, not evidence.

The agency talks about AI search using vague buzzword phrasing without explaining the actual retrieval mechanics or showing a citation report. If you cannot tell what the agency would actually do for your business in concrete terms, the AI search positioning is decoration, not capability.

The agency cannot articulate when SEO is the wrong investment. Every channel has a wrong fit. An agency that pretends SEO is universal for every business is either undiscriminating or selling.

The agency dodges contract questions. Month-to-month or three-month terms with clear exit language is the buyer-friendly norm. Twelve-month lock-ins without performance clauses favor the agency, not you.

The agency uses "we are different" as the differentiator. Every agency claims that. Specifics are the only credible answer (proprietary processes, named team members, verifiable client outcomes, platform partnerships).

How to compare proposals

Most buyers compare SEO proposals on retainer price alone, which is the worst possible comparison. The four-dimension framework below produces better decisions.

Compare scope at the line-item level. Get every proposal to break out hours or deliverables across strategy, technical SEO, content production, link building, AI search tracking, digital PR, and reporting. The cheapest retainer often disappears when you see the line items, because half the scope you expected is missing.

Compare team seniority and account load. A $5,000 retainer staffed by a senior strategist with eight accounts almost always outperforms a $3,000 retainer staffed by a junior coordinator with 25 accounts. Cheaper hourly rates lose to fewer, higher-quality hours.

Compare measurement clarity. Ask each agency to define the three metrics they will report monthly and the threshold at which they will recommend strategy changes. Agencies that cannot define their own measurement bar tend to drift, and the drift shows up in your quarterly numbers six months too late.

Compare AI search capability honestly. Treat any proposal that does not include explicit AI surface tracking and citation strategy as incomplete. Most SEO retainers in 2026 still do not include this work by default, which means you either pay the agency extra for it, hire a specialist on the side, or build it in-house. None of those options is free, but the cheapest path is usually a single agency that does both. Our comparison of in-house SEO versus agency and our breakdown of AEO/GEO tracking tools for B2B SaaS cover that math in more depth.

Ask the questions before the budget gets harder

The buyer who runs this checklist on three or four shortlisted agencies will leave the process with a clear winner and a clear reason for the choice. The buyer who skips the checklist will usually leave with the agency that pitched best, which is rarely the agency that delivers best. The 23 questions above were built specifically so that the answers separate strong agencies from weak ones in 20 minutes per call, not five hours per evaluation.

If the next agency conversation on your calendar is a real one, get a current-state baseline first so you can interrogate proposals against actual performance, not abstract promises. Look at how Passionfruit's SEO service and GEO service approach the audit and citation tracking work, see the cross-platform measurement inside Passionfruit Labs, and talk to the team before the budget conversation gets harder.

Frequently asked questions

The questions below come up most often when marketing leaders and founders are sitting down to vet SEO agencies in 2026.

What are the top 5 questions to ask an SEO agency before hiring?

The top five are: (1) Walk me through the first 90 days for our business specifically. (2) Show me a real client report from the last quarter, including both Google rankings and AI citation share. (3) How do you measure success beyond keyword rankings and traffic? (4) Which AI surfaces (ChatGPT, Perplexity, Gemini, AI Overviews, Claude) do you optimize for, and how is that work different from classic SEO? (5) Who will work on our account, what are their seniority levels, and how many other accounts do they handle?

Should an SEO agency be doing GEO and AEO in 2026?

Yes. AI Overviews now appear on roughly 48% of Google queries, and ChatGPT, Perplexity, and other AI surfaces handle a growing share of category-defining searches. An SEO agency that does not include AEO and GEO work is optimizing for half the surface area buyers actually search across. The right answer to the question is "yes, here is how we do it, here is a real citation report, here is how we measure it."

How can I tell if an agency really does AI search work or is just using it as a marketing label?

Ask for a citation report from a current client showing share across at least three AI surfaces with quarter-over-quarter deltas and competitor benchmarks. Ask which tracking platform they use (Passionfruit Labs, Profound, Otterly, or a structured manual process). Ask how they handle Reddit, YouTube, Wikipedia, and review-site presence, since AI surfaces cite those heavily. If the agency cannot answer those three questions concretely, the AI search positioning is marketing, not capability.

What is a fair monthly retainer for SEO in 2026?

Most US small and mid-market retainers in 2026 land between $1,500 and $5,000 per month, with the survey-reported average around $3,200 per month. Mid-market and enterprise engagements run $5,000 to $15,000 per month. Highly competitive verticals and enterprise programs cross $15,000. Compare scope and team seniority, not just price.

What should an SEO agency contract look like?

A buyer-friendly SEO contract has month-to-month or three-month terms, no auto-renewal lock-ins, written cancellation language, clearly defined deliverables, an explicit definition of what assets transfer back to you on exit (analytics access, content drafts, link inventories, dashboards), and a performance clause that lets you exit if defined metrics are missed for two consecutive quarters.

How long should I expect to wait for results from an SEO agency?

Most legitimate engagements show early ranking movement and content velocity in 60 to 90 days, measurable traffic gains in the 4 to 6 month range, and meaningful revenue or lead-flow impact at 6 to 12 months. AI citation share can move faster (some clients see early shifts in 4 to 8 weeks) because citation depends on content structure and freshness more than backlink accumulation. Anyone promising significant results in 30 days is selling, not delivering.

How many SEO agencies should I get proposals from?

Three to four is the right number for a mid-market engagement. Fewer than three and you have no comparison. More than five and you waste internal time without improving the decision. Brief each shortlisted agency on the same goals, the same baseline data, and the same evaluation criteria so the proposals are comparable.

What is the most common mistake buyers make when hiring an SEO agency?

The most common mistake is comparing proposals on retainer price alone instead of on scope, team seniority, measurement clarity, and AI search capability. The cheapest retainer is rarely the cheapest outcome over 12 months, because the gap usually shows up in missing work, weaker reporting, or no AI search coverage that you end up paying to add separately later.

What if my budget is below $2,500 per month?

Below $2,500 per month, the realistic options are a freelance senior SEO consultant on a fractional retainer, a specialty agency that focuses on one part of the discipline (technical SEO, content production, or AI search visibility), or a hybrid model with one part-time in-house lead supported by project-based agency work. Full-service SEO agencies tend to under-deliver below $2,500 because the headcount cost to produce real work does not fit the budget.