What a credible AEO case study contains (and what a hollow one looks like)

TL;DR:

• A credible Answer Engine Optimization (AEO) case study names the client, shows pre-campaign baselines, documents the methodology, and traces citations to CRM-verified pipeline, with caveats stated.
• Proof pattern:
  • incident.io: AI visibility 38% to 64%, +22% organic meetings booked
  • Sova Assessment: +167% organic demo requests, organic became the #1 channel
  • Gladia: 7x sales-accepted leads in 4 months, 93% from LLM search
  • Granola: +4,600% monthly activated users from AI search, +134% organic sessions, +558% non-branded AI citation rate
• Red flags: anonymous clients, round numbers, no timeframe, no attribution path, no stated limitations.
• Action: run the six-point audit below on every agency case study before you sign a retainer.

Most agencies write their case studies to close a sale, not to survive a board review. That problem gets expensive in AI search, where agencies can produce visibility scores and citation growth charts easily but you cannot verify them without CRM data. If you are shortlisting AEO agencies right now, every AEO case study in front of you deserves the same scrutiny you would apply to any pipeline claim: baseline, intervention, timeframe, and a path to revenue. This guide gives you the six data points a credible case study contains, the red flags that mark a hollow one, and the exact numbers we publish for our own clients at Discovered Labs.

A credible AEO case study names the client, shows pre-campaign baselines for AI visibility and pipeline, documents the methodology (like our CITABLE framework), sets a realistic timeframe (initial citations in 1-2 weeks, 40% citation rate by month 4), traces citations to CRM-tracked revenue (MQLs, demos, closed-won deals), and states attribution limitations honestly. A hollow one hides behind anonymous logos, reports citation lifts without baselines or buyer-intent query context, and stops at visibility scores rather than connecting to Salesforce or HubSpot pipeline. Treat each case study like an investment memo: if the numbers cannot survive your CFO's questions, it is sales material, not evidence.

Why case study credibility matters for CMOs evaluating AEO agencies#

Your CFO does not care about citation rate. They care whether marketing-sourced pipeline justifies the retainer, and GA4, HubSpot, and Salesforce already give you three different answers to the same attribution question. An AEO case study that stops at "visibility up" hands you a number you cannot defend in a quarterly board review, which is why we published a guide on proving AI search ROI in terms a finance team accepts.

The deeper problem is mechanical. Google crawls, indexes, and ranks documents, while LLMs retrieve passages and synthesize a single answer. Traditional SEO and answer engine optimization share foundations but differ in retrieval technology, which is why our GEO audit vs SEO audit comparison shows what traditional audits miss. Because the retrieval technology differs, the proof standard differs too.

Research documents the performance gap between the two retrieval systems. Dense retrievers beat sparse methods like BM25 by 9 to 19 points on top-20 passage retrieval accuracy in the original Karpukhin et al. paper, because dense passage retrieval matches meaning rather than exact keywords while BM25 matches lexical terms and term frequency.

Ahrefs data from early 2026 shows about 38% of AI Overview citations came from pages ranking in the top 10, so most of what AI cites is not what ranks. Treat that as a snapshot, not a precise trend, because Ahrefs improved its citation parsing between studies. Even so, a strong Google position is weak evidence of AI visibility on its own.

Tracking AI impact in your CRM#

You prove pipeline in your CRM, not in a dashboard screenshot. The mechanics are straightforward: you configure UTM parameters in Salesforce to preserve lead source data, and hidden form fields carry UTM data into HubSpot alongside the lead's name and email. We also add a "How did you hear about us" field during onboarding, because click-based tracking files most AI-referred visits as Direct or Organic, and self-reported answers catch what UTMs miss, like a prospect who asked Claude for vendor options and then typed your URL directly.

If an agency cannot describe this setup in detail on the first call, they cannot trace citations to revenue for you. Some agencies pitch a proprietary attribution layer instead: Omnius, for example, builds measurement into its platform, and our approach maps AI-referred leads directly into your existing HubSpot or Salesforce pipeline, so your CFO reads one set of numbers instead of two.

Why vanity metrics obscure real pipeline#

A citation on a low-intent query adds nothing to your pipeline. If an agency reports a 50% citation rate lift but cannot name the queries, you have a vanity metric, because the queries that matter are buyer-intent queries like "best incident management software," not "what is incident management."

Off-page signals matter just as much. Our analysis of 144,000 AI citations found Reddit referenced in roughly 27% of ChatGPT search results while appearing in under 1% of visible citations, which is why we treat Reddit as citation infrastructure and run managed Reddit programs for clients. A case study that never mentions third-party consistency describes half the mechanism.

6 valid data points every AEO case study needs#

Six elements separate evidence from sales material. If a case study misses any of them, discount everything else it claims.

Named client with verifiable footprint#

Anonymous case studies (like "a leading fintech") prevent you from verifying anything. A named client lets you check the company's footprint, test the queries yourself, and ask to speak with the marketing leader on record. We publish named clients with named contacts. If an agency claims an NDA prevents client naming, ask for the verification steps we describe in the anonymized-results section below. Case studies are only one of four review sources, and they are the most commercially biased. We weight all four in where AEO agency reviews live.

Establishing pre-campaign performance benchmarks#

Without a baseline, every lift is unverifiable. A credible case study documents starting citation rate, mention rate, and share of voice across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews, plus organic sessions and pipeline. You can measure your own baseline with our free AI visibility audit checklist and compare it against published AEO benchmarks. Score individual pages with the AEO content evaluator to see where extractability falls short. incident.io started at 38% AI visibility before our engagement, which is what makes the move to 64% meaningful.

Documented execution of AEO methods#

"Secret sauce" is a red flag. Credible agencies name their methodology, and ours is the CITABLE framework: Clear entity and structure, Intent architecture, Third-party validation, Answer grounding, Block-structured for RAG, Latest and consistent, and Entity graph and schema. RAG (retrieval-augmented generation) lets the model look up passages before answering, like an open-book exam, which is why our CITABLE methodology structures every section to answer one question independently.

LLMs also cite claims that appear consistently across independent sources, so third-party validation carries weight that on-page work alone cannot, and entity SEO makes your brand recognizable to models in the first place. Editorial quality still matters: Omniscient Digital produces strong B2B editorial content, and our Omniscient Digital comparison shows how CITABLE structures that caliber of content for passage retrieval so the same asset wins rankings and citations.

Milestones for AEO performance gains#

Credible case studies commit to a timeline. We commit to initial citations in 1 to 2 weeks, meaningful citation rate lift to roughly 40% on priority buyer queries by month 4, and full optimization across web search, citations, and training data by month 6. A managed AI visibility audit sets the baseline these milestones hang on, and an AEO audit defines the priority query set. Any agency promising full visibility in 30 days is describing their sales process, not their delivery.

Tracing AI citations to closed revenue#

A credible case study walks you through five steps from citation to revenue:

1. Citation: A buyer asks ChatGPT or Claude a category question and the client is cited as a source.
2. Visit: The buyer clicks the citation link or types the URL directly.
3. Capture: The demo form carries UTM parameters and a self-reported "How did you hear about us" answer into HubSpot or Salesforce.
4. Conversion: The lead becomes an MQL, then an opportunity, then closed-won, with the source field intact.
5. Reporting: AI-referred pipeline appears in the monthly board slide next to every other channel.

We report MQLs, demos, and pipeline rather than citation counts alone. If an agency's case study stops at step one, the other four steps are where their proof falls apart.

Clarifying attribution limitations#

AI search attribution is probabilistic, and anyone claiming otherwise is selling certainty they do not have. Citation tracking tools estimate visibility by sampling prompts, and our AI tracking platforms test shows how design flaws in those tools distort the numbers. Click-based attribution misses most AI-referred visits, self-reported fields depend on buyer memory, and GA4 will never reconcile perfectly with Salesforce. A credible case study states these limits upfront. If an agency claims perfect attribution, they have just told you they do not understand the measurement problem.

Evidence in action: four AEO success stories#

Every element above is checkable against our own published results. Here are the four engagements we point to when CMOs ask what credible looks like, with the numbers exactly as we publish them. All four are documented in full, with client quotes and independent ratings, on our Discovered Labs reviews page.

incident.io: linking AI visibility to pipeline#

incident.io came to us with strong Google rankings and weak AI visibility. We restructured priority content with the CITABLE framework and built third-party consistency across the sources LLMs trust. AI visibility moved from 38% to 64%, closing the gap on PagerDuty, and organic meetings booked grew +22%. Their CMO, Tom Wentworth, is on the record:

"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 baseline (38%), the endpoint (64%), and the pipeline metric (+22% meetings) are all stated, which is what makes the claim auditable.

Sova Assessment: +167% organic demo requests, #1 lead channel#

Sova Assessment needed organic to carry more pipeline weight. We ran SEO and AI search as one operation: technical fixes, CITABLE-structured content, and off-page consistency work. Organic demo requests grew +167%, and organic search became the #1 channel for leads and MQLs, with Sabina Reghellin, their Content Marketing Manager, as the named contact on the engagement. The point for your board: we state the result as channel contribution, not an abstract visibility score.

Gladia: 7x sales-accepted leads, 93% from LLM search#

Gladia gave us four months. We mapped priority buyer queries, shipped CITABLE content against them, and built the third-party footprint. Sales-accepted leads grew 7x in 4 months, and 93% of AI-referred leads came from LLM search. That 93% figure matters: it shows we instrumented the attribution path well enough to say where leads came from, not just that they arrived.

Granola: +4,600% monthly activated users from AI search#

Granola is an AI notepad competing in a crowded productivity category. We ran one program across content, technical SEO, and links and mentions. Monthly activated users attributed to AI search grew 4,600% from a January 2026 baseline, organic sessions grew 134%, and the non-branded AI citation rate grew 558%. The baseline date is what makes this auditable: a percentage without a start point is the exact red flag this guide warns about, so we publish ours. Julien, Growth Marketing at Granola, is the named contact on the engagement.

Identifying hollow claims in agency reporting#

Hollow case studies share five failure modes. Each one maps to a missing data point from the six above.

Why ROI claims fail without baselines#

"300% growth in citations" sounds strong until you learn the client went from 2 citations to 8. The math is technically correct and completely misleading, which is the point: agencies have used percentages without denominators as their oldest reporting trick. Ask for the absolute baseline, the query set size, and the timeframe. If any of the three is missing, the percentage is decoration.

Missing pipeline and attribution data#

If the last mile of a case study is a visibility dashboard screenshot, it proves nothing about revenue. Impressions, visibility scores, and citation counts are leading indicators at best. The case study should end in the CRM: MQLs, demos, opportunities, or closed-won deals with a described source path.

Missing proof of client results#

You should treat self-reported agency metrics as marketing copy until a client puts their name on them. On-record testimonials let you verify the engagement actually happened and call the reference. Here is the standard, from Liam Collins, Head of Performance at Instantly:

"We're working with Discovered Labs and have achieved excellent results. Highly recommend the team's delivery..." - Liam Collins, Head of Performance at Instantly

You can verify a named person with a named title at a named company. You cannot verify a logo cloud.

Missing context on attribution math#

Some agencies claim credit for demand that already existed. If a client had strong branded search before the engagement, the agency may be claiming credit for brand traffic they never influenced. Ask how the agency separates baseline brand demand from campaign-driven demand, and what the client's non-brand pipeline did during the same period.

Why traffic growth ignores revenue#

Traditional SEO agencies often optimize for high-volume, low-intent keywords because those keywords inflate sessions cheaply. The result is a blog that outranks the money pages and traffic that never converts. Start with buyer-intent queries for exactly this reason. A credible case study reports demos and MQLs against buyer-intent queries, not session counts against informational ones.

Assessing AEO credibility: a CMO checklist#

Bring this checklist to every agency evaluation call. You can complete it in ten minutes, and it will filter out agencies that cannot prove pipeline impact.

The team composition is a direct reflection of our approach. Our CTO Ben Moore, an ex-Stanford AI researcher who built fraud systems at Stripe and Coinbase, works alongside SEO strategists and editors to run retrieval optimization and traditional SEO as one connected operation. Our CEO Liam Dunne scaled a B2B SaaS past $20M ARR from the marketing seat, so the reporting standard comes from someone who has defended a marketing budget at board level.

Key questions for assessing impact#

Ask these questions on every call and note which agencies answer with specifics versus adjectives:

1. CRM proof: Can you show me how a specific citation converted to an opportunity in your client's CRM?
2. Baseline: What was the starting citation rate, on which query set, measured with which tool?
3. Query quality: Which buyer-intent queries drove the pipeline you claim, and what did non-brand pipeline do over the same period?
4. Reference: Who is the named contact on this engagement, and can I call them?
5. Method: What did you actually ship, meaning which pages, which framework, and which off-page placements?
6. Limits: What can your measurement not capture? An agency that answers "nothing" fails.

Six questions, ten minutes, and most weak case studies collapse by question three.

Red flags in agency performance data#

Use this table when you read any agency's proof page:

We hold ourselves to the same standard, which is why we publish a ranked comparison of AEO agencies with named competitors and named trade-offs.

Linking AI traffic to revenue#

The technical requirements are boring and non-negotiable: UTM conventions on every URL you control, a self-reported attribution field on every form, and source fields that survive the trip from form to CRM to board slide. Our AI search audit checks this plumbing before any content ships, and we run one connected operation across both search surfaces so the data thread never breaks. This is also where the scope of the engagement matters: Quoleady focuses on B2B SaaS content, and we combine content operations with technical SEO and Reddit marketing because LLMs reward consistent claims across independent sources, not volume alone.

Beyond vanity metrics: assessing AEO performance#

Four closing questions separate durable proof from one good quarter.

Evaluating anonymized agency results#

Strict NDAs are real, especially in security and regulated categories, so an anonymized case study is not automatically hollow. The test is specificity. Our own NDA-protected engagement reports 6x AI-referred trials in 7 weeks and 3,500+ AI-referred trials: unrounded numbers, a defined timeframe, a named metric. "A B2B SaaS company grew trials 6x" with no timeframe and no absolute figures is marketing copy. Ask for anonymized CRM screenshots and a methodology walkthrough before you accept or discount the claim.

The pipeline gap in AEO case studies#

Most AEO programs fail between traffic and pipeline. Content earns citations, citations earn visits, and then weak forms, missing UTMs, and unmapped source fields lose the thread. Credible agencies report the full funnel because they instrumented it from day one. If a case study shows traffic without pipeline, assume the gap was never closed.

How long should results take in a credible AEO case study?#

A credible case study covers at least a full quarter. Expect initial citations in 1 to 2 weeks, meaningful citation rate lift by month 3 or 4, and full optimization across web search, citations, and training data by month 6. Anything promising faster is describing sales copy, not delivery.

Verifying agency-reported performance metrics#

You can verify most claims yourself: run the agency's claimed queries in ChatGPT and Perplexity, check the client's footprint, and compare reported numbers against independent AI visibility audit tools. The full audit protocol, including the four questions that expose a rebadged SEO retainer, is in how to verify an AEO agency's claimed AI visibility results. Better yet, establish your own baseline before you sign anything.

Get your baseline before you sign#

Request a Search Visibility Diagnostic: €4,370 one-off, 100% credited to your first retainer if you sign within 14 days. You get an objective audit of your AI visibility across ChatGPT, Claude, Perplexity, and Gemini, plus your current Google baseline, so every agency claim gets measured against your numbers, not theirs. Or book a call and we'll tell you honestly whether we're a fit.

FAQs#

How long does it take to see results in an AEO campaign?#

Initial citations typically appear within 1 to 2 weeks, while a meaningful citation rate lift toward 40% on priority buyer queries takes 3 to 4 months of consistent optimization. Full optimization across web search, citations, and training data takes about 6 months.

What is the cost of Discovered Labs' AEO and SEO services?#

Our Establish package is €7,995 per month, month-to-month (€6,995 per month on the 6-month rate), and Compete is €12,995 per month (€10,995 on the 6-month rate). Every tier is on public pricing with no annual lock-in.

How do you track AI-referred leads in HubSpot or Salesforce?#

We configure UTM parameters on AI-referred URLs and add a self-reported "How did you hear about us" field to your forms, then map both to lead source fields. AI-sourced MQLs flow directly into your existing pipeline reports.

Key terms glossary#

Citation rate: The percentage of AI-generated answers for a defined query set that explicitly cite your brand as a source.

Passage retrieval: The process where LLMs pull semantically relevant text blocks to synthesize an answer, rather than ranking entire documents.

Information consistency: The alignment of accurate claims about your brand across your site, Reddit, and independent publications, which LLMs use to verify what to cite.