How to verify an AEO agency's claimed AI visibility results

TL;DR:

• Demand raw proof: Ask for prompt logs, dated screenshots, and the exact query set behind any claimed citation rate. Aggregated visibility scores hide cherry-picking.
• Prioritize citation rate over mention rate: A citation grounds the answer in your content, while a mention only names you. Measure both against a fixed query set.
• Pair visibility with pipeline: incident.io's AI visibility rising from 38% to 64% alongside +22% organic meetings booked shows visibility tied to business outcomes, which is the standard to demand from any agency.
• Audit attribution yourself: Combine UTM tagging, AI-referral tracking in GA4, CRM fields, and a self-reported "how did you hear about us" field.

Verifying organic search results used to mean checking rankings and traffic. Now buyers research inside ChatGPT, Claude, and Perplexity before visiting your site, and Ahrefs data from early 2026 shows only about 38% of Google AI Overview citations came from pages ranking in the top 10. A ranking report tells you nothing about AI visibility, which is why knowing how to verify AEO agency results is now a core procurement skill for B2B SaaS marketing leaders.

This guide gives you a vendor-agnostic audit protocol: the proof to demand, the metrics that hold up in a board review, the attribution setup that connects AI visibility to pipeline, and the red flags that expose a rebadged SEO agency.

If you are still at the shortlist stage, start with where AEO agency reviews live and what each source actually tells you.

To verify AEO agency results, demand raw prompt logs, dated screenshots of LLM responses, and a clear attribution path to pipeline. Avoid agencies that report only aggregated visibility scores without disclosing the underlying queries. A reliable verification process pairs citation rate lifts on a fixed set of buyer-intent queries with downstream business metrics like MQL growth and demo requests. Audit the agency's tracking methodology and set up custom CRM fields, and you will know within one quarter whether your AEO investment produces measurable pipeline.

The hidden challenges of validating AI results#

Validating AI search results is harder than validating rankings for one reason: the output is probabilistic. Ask ChatGPT the same question twice and you can get two different answers with two different sets of citations. Any verification method has to account for that variance instead of pretending it does not exist.

Mapping the AI research path to revenue#

A buyer can ask ChatGPT for a shortlist, read three synthesized answers, and visit your site days later through a branded search or a direct visit. GA4 files that session as direct or organic, and the AI touchpoint disappears from your attribution model.

This is the attribution gap, and any agency that claims to have closed it completely is overstating its measurement. The practical response is triangulation: combine AI-referral data, self-reported attribution, and citation tracking to bound the true number. We lay out that approach in our guide to auditing AI search visibility.

Beyond clicks: measuring AI growth#

Zero-click evaluation is now normal buyer behavior, so traditional CTR and impressions cannot carry your reporting. The KPI set that matters includes citation rate, mention rate, share of voice, AI-referred sessions, and qualified pipeline attribution.

Treat every AI visibility tool as an estimate: our test of AI tracking platforms shows how measurement methodology distorts what these platforms report. The invisible influence problem also runs deeper than tools. Our analysis of 144,000 AI citations found Reddit referenced in roughly 27% of ChatGPT search results, a signal that never appears in a ranking report.

Red flags that signal a rebadged SEO agency#

Google scores documents and returns a ranked list. LLMs retrieve semantically relevant passages and synthesize one answer, and dense retrievers outperformed BM25 by 9 to 19 points on passage retrieval in the original DPR paper. The retrieval technology is genuinely different.

Therefore, the signals differ too. Google weighs links and rankings, while LLMs weigh passage structure, entity clarity, and third-party consistency. Google's AGREE research shows LLMs self-ground claims against retrieved documents, which rewards claims repeated consistently across independent sources. An agency pitching backlinks and meta descriptions as its AEO offer is selling SEO with a new label. To understand the real mechanical differences, see our breakdown of GEO audit versus SEO audit, entity SEO for AI, and this video on SEO versus AEO.

Critical questions to surface agency red flags#

Four questions, asked early in the RFP process, separate evidence-based agencies from relabeled SEO retainers.

Demand proof of AI citation sources#

Ask: "Can you show me the exact prompt logs and dated screenshots behind these numbers?" Raw logs include the prompt text, the platform, the date, and the full response. Aggregated visibility scores without the underlying queries let an agency cherry-pick wins.

Also ask which tracking tool they use and how it samples responses, because AI visibility audit tools vary widely in methodology. Visible citations understate real influence: Reddit appeared in only 0.35% of visible citations in our dataset despite being referenced in roughly 27% of ChatGPT query processing, so screenshot evidence alone never tells the whole story.

Why citation rates outperform mentions#

Citation rate is the percentage of priority buyer queries where an LLM cites your content as a source. Mention rate is the percentage of answers that name your brand, with or without a source link.

Citation rate is the stronger signal. A mention means the model knows your name, possibly from training data with no way for a buyer to click through. A citation means the model grounded its answer in your content, with a source you can audit. Both metrics need a fixed query set to mean anything: an agency can inflate either one by testing low-value queries where you already appear, but citation rate on 50 pre-agreed buyer-intent queries resists gaming. Our AEO benchmarks guide covers query-set design in detail.

Tracing AI sources to revenue#

Ask: "How do you connect these citations to our HubSpot or Salesforce pipeline?" A credible answer names specific mechanisms: UTM conventions on links the agency controls, referral monitoring for chatgpt.com, perplexity.ai, and claude.ai in GA4, and a self-reported attribution field on your demo form.

If the answer stops at "visibility," keep looking. Visibility without an attribution path gives you a chart, not a business case.

Check real case studies for pipeline impact#

The standard to demand is a visibility metric paired with a business metric, published with enough detail to audit.

The same figures, with client testimonials and independent ratings alongside them, are consolidated on our reviews page. The full list of what a case study needs to clear that bar is in what a credible AEO case study contains.

Use this pattern as your reference when evaluating vendor case studies. We published every figure in this table on our case studies page.

We documented the anonymized engagement in a B2B SaaS AEO case study with a recorded case study walkthrough. Sova Assessment saw +167% organic demo requests, and organic search became its number one channel for leads and MQLs.

How to audit agency claims for accuracy#

Run this protocol during the RFP stage, before any contract is signed. We walk through the underlying mechanics in our AI search guide for 2026.

Once the contract is signed, the benchmark shifts to delivery. We set out the checkpoint-by-checkpoint expectations in what results to expect in the first 90 days.

Correlating AI metrics with MQL growth#

Look for directional correlation over a quarter: as citation rate on your priority query set rises, AI-referral sessions and self-reported AI leads should rise with it.

Auditing AI lead attribution logic#

Ask the agency to walk you through its attribution math, then verify the setup in your own stack:

1. Add a self-reported field: Put "how did you hear about us" on demo and contact forms, with ChatGPT, Claude, Perplexity, and AI search as options.
2. Create a CRM property: Map a custom field in HubSpot or Salesforce to capture the answer and carry it to the opportunity record.
3. Monitor LLM referrers: Build a GA4 segment to track sessions from chatgpt.com, perplexity.ai, claude.ai, and gemini.google.com. Configure referral exclusions if needed to preserve attribution.
4. Standardize UTM tagging: Tag every link you control that AI platforms might cite.
5. Reconcile monthly: Compare referral data, self-reported answers, and CRM source fields, then report the range honestly.

This is the reporting frame we describe in proving AI search ROI to finance leadership.

Mapping your baseline visibility metrics#

Verification starts before the engagement does. Run 50 priority buyer queries across ChatGPT, Claude, Perplexity, and Gemini, then record where you are cited, mentioned, or absent. Our free AI visibility audit checklist covers the DIY version, and our breakdown of what an AEO audit covers shows the professional scope. You can also score individual pages with the free AEO content evaluator.

Set timeline expectations from our published methodology: initial citation movement in 2 to 4 weeks, meaningful share-of-voice gains in 6 to 8 weeks, and compounding effects over 3 to 6 months. An agency promising faster is selling a guarantee it cannot keep.

Case study: measuring incident.io citation growth#

Here is what verifiable proof looks like, using our own published data as the worked example.

Validating 26-point growth in AI visibility#

incident.io's AI visibility rose from 38% to 64% over four months, closing the gap on PagerDuty in AI share of voice. The full detail is in the incident.io case study.

We measured the 26-point lift in absolute terms against a fixed set of buyer-intent queries. That fixed query set is what makes the number auditable, because no one moved the goalposts between the baseline and the result.

Validating the 22% meeting uplift#

The visibility lift came paired with business metrics:

• +22% organic meetings booked: Including a 50% month-over-month jump from January to February.
• Top-3 keyword rankings: Expanded from 142 to 228 over the same period, showing both surfaces moved together from one content operation.

That pairing of a visibility number with a pipeline number is the exact pattern to demand from any agency claiming AI visibility results.

Validating the attribution logic#

We track the pipeline connection through self-reported attribution and CRM source data, and we do not attribute any closed deal to a single citation. That honesty is part of the proof.

"Before Discovered Labs, we were using homegrown LLM prompts, without a clear strategy for what to optimize for or exactly how best to structure content." - Tom Wentworth, CMO at incident.io

The takeaway for your scorecard: demand the pairing of visibility and pipeline, and accept that attribution will be probabilistic rather than perfect.

Audit protocol: the Gladia results breakdown#

Validating AI-referred lead claims#

Gladia grew sales-accepted leads 7x in 4 months, with 93% of AI-referred leads coming from LLM search. We published both figures in full on our case studies page. The key is that "AI-referred leads" is a defined, tracked segment in the CRM, which makes the 93% figure auditable rather than anecdotal.

An agency that cannot define its lead segments cannot verify its claims.

How Gladia tracks AI referrals#

The tracking behind those numbers combines three mechanisms we set up as standard: UTM tagging on links we control, referral monitoring for LLM domains in GA4, and a self-reported attribution field on demo forms feeding the CRM.

That setup puts AI-sourced leads into the same MQL-to-opportunity reporting as every other channel. It is the attribution path to demand from any vendor: a stated mechanism you can inspect, not a black box labeled "AI influence."

6 essential criteria for your AEO vendor scorecard#

Score each criterion on evidence rather than promises, and weight the scorecard toward pipeline proof.

Insist on transparent performance data#

Score vendors on whether they provide raw prompt logs, dated screenshots, screen recordings, and the full query list behind any claimed citation rate. Refusal to share raw data signals the agency cannot defend its numbers, which is a disqualifying answer, not a negotiation point.

Establish a monthly performance review#

Score vendors on reporting cadence and substance. You want a monthly narrative that connects citation rate, AI-referred sessions, and pipeline, not a data dump of screenshots. Ask to see a sample report before signing.

Integrating AI leads into Salesforce#

Score vendors on whether they will configure the tracking inside your stack: lead-source values, custom fields, and the reports your board will actually see. An agency that cannot touch your CRM cannot prove pipeline impact.

Distinguishing vanity from revenue metrics#

Score vendors on their metric hierarchy. Citation rate, share of voice, and AI-referred MQLs belong at the top, while aggregate visibility scores and "impressions" belong at the bottom. We built our CITABLE framework with a team that includes our CTO Ben Moore, an ex-Stanford AI researcher. That kind of ML depth signals the methodology criterion is met.

Use these weights as a starting point, and adjust the percentages to match your board's priorities:

Pricing transparency and flexibility#

Score vendors on whether pricing is public, terms are month-to-month, and there is no annual lock-in requirement. Agencies that hide pricing until a sales call are optimizing for closing deals, not for client fit. Transparent pricing signals the agency is confident its results justify the fee, and a month-to-month structure puts the performance obligation on the right party.

Specialization#

Score vendors on whether organic search is their primary focus, or whether AEO is one service in a generalist agency selling paid, social, and creative in the same breath. A shop whose core practice is organic search thinks differently about content, retrieval signals, and measurement than one running AEO as an upsell. Ask what percentage of their active retainers are organic search engagements.

Red flags when vetting AI visibility partners#

These are the SEO agency red flags that apply whether the pitch is labeled SEO, AEO, or GEO. The same audit logic applies when you verify SEO agency results, though the proof points differ by surface. Any one of these flags should trigger a harder question, and two or more should end the conversation.

Missing attribution metrics for results#

The agency cannot explain how it tracks or attributes AI-referred traffic. A real measurement plan names the tools, the CRM fields, the UTM conventions, and the reporting cadence. If the answer is "you will see it in the visibility score," there is no measurement plan.

Only visibility lifts, no pipeline tie-back#

The agency promises "10x visibility" but refuses to connect its retainer to pipeline-adjacent metrics like MQLs, demos, or meetings. Visibility is an input, not an outcome.

Missing evidence for performance claims#

The agency cites industry stats you cannot trace and has no named, published client case studies with business metrics. Unverifiable claims dressed as proof are worse than no claims at all.

Inability to track AI-sourced MQLs#

The agency cannot set up or audit the CRM workflows that capture AI-referred leads. Without that plumbing, you will never build the board slide that defends the spend.

The agencies that welcome this kind of scrutiny are the ones worth hiring. If you want a verified baseline before your next vendor decision, our Search Visibility Diagnostic (€4,370 one-off, $4,995 USD) audits your brand across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews. The fee is credited in full to your first retainer if you sign within 14 days. You can see what a managed audit delivers, explore our AEO and SEO services, or watch our full AI search guide before you book a call. We will tell you honestly if we are a fit.

FAQs#

What is the difference between citation rate and mention rate?#

Citation rate measures the percentage of priority buyer queries where an LLM cites your content as a source, while mention rate measures the percentage of answers that name your brand, with or without a source. Weight citation rate higher, because it is auditable and closer to a site visit.

How long does it take to see measurable AI visibility results?#

Initial citation movement typically appears within 2 to 4 weeks of publishing optimized content. Meaningful share-of-voice gains follow in 6 to 8 weeks, with compounding effects building over 3 to 6 months, and we target roughly 40% citation rate on priority queries within that window.

Can a closed deal be attributed to a specific ChatGPT or Claude citation?#

No, and we do not claim any closed deal can be attributed to a specific Claude or ChatGPT citation, because that data is not published or verifiable. Track overall AI-referred pipeline and self-reported attribution instead.

How much does a professional AI visibility audit cost?#

Our Search Visibility Diagnostic costs €4,370 one-off ($4,995 USD), covering a full audit across major LLMs, an entity map, and a schema review. The fee is credited in full to your first retainer if you sign within 14 days.

Where can I see Discovered Labs' full case study data?#

You can view verified, unrounded data for Gladia, incident.io, and Sova Assessment on our case studies page, including how organic became the number one channel for leads and MQLs for Sova.

Key terms glossary#

Citation rate: The percentage of priority buyer queries where an LLM cites your content as a source. Calculated as cited queries divided by total queries tested.

Mention rate: The percentage of AI-generated answers that name your brand, with or without a source citation. A mention shows awareness, while a citation shows source-level trust.

Passage retrieval: The process by which an LLM finds and extracts semantically relevant blocks of text to synthesize an answer. It replaces document-level ranking, which is why SEO metrics alone cannot verify AI visibility.

Information consistency: The alignment of accurate claims about your brand across independent sources like Reddit, industry publications, and your own site. LLMs use it as a trust signal when choosing citations.

CITABLE framework: Discovered Labs' 7-component content framework built for LLM retrieval: clear entity and structure, intent architecture, third-party validation, answer grounding, block-structured for RAG, latest and consistent, and entity graph and schema.