Where AEO agency reviews live and what each source actually tells you

TL;DR

• Public directories are a starting point, not a decision. Clutch verifies that projects happened, but its category rankings favor paid profiles in visibility.
• G2 is built for software, so agency profiles are sparse and low-signal.
• Vendor case studies count only when they show exact metrics, the technical approach, and an attribution path.
• Private peer referrals carry the highest signal because no marketing filter sits between you and the truth.
• In our view, peer referrals should carry the most weight in your scorecard, followed by verified case studies, then Clutch, and finally G2 plus unverified testimonials.

Most agency review directories are commercial platforms in disguise. If you shortlist an organic search partner on aggregate star ratings alone, you're buying their marketing budget, not their technical capability. Answer Engine Optimization (AEO) agency reviews typically appear in four places: Clutch, G2, private peer referrals, and vendor-hosted case studies, and each source carries a different mix of verification rigor, commercial bias, and pipeline signal. The partner you pick has to perform across web search, AI citations, and model training data, so the reviews you trust need to speak to all three. This guide breaks down what each source actually tells you, how to filter the noise, and how to weight the inputs so your shortlist decision holds up in front of your CFO.

AEO agency reviews are not equally reliable across sources. Clutch runs verified client interviews, but its category rankings give paid profiles greater visibility, so position can reflect ad spend alongside delivery quality. G2's platform structure makes agency profiles tend to be sparse and lower-signal than software product reviews. Self-published testimonials show best-case outcomes by design. Private peer referrals carry the highest pipeline signal because they skip every marketing filter. The reliable pattern: weight peer referrals and verified pipeline data first, treat directory scores as baseline verification, and ask every shortlisted agency to prove impact in MQLs, demos, and revenue.

How to filter noise from review authority#

The CMOs who get agency selection right treat reviews as data to be audited, not scores to be compared.

Separating verified results from sales noise#

The line between signal and noise is quantitative verification. "They were great to work with" tells you nothing. "Organic demo requests up 167%" is something you can audit. LLM retrieval differs from Google ranking, so a review praising keyword rankings alone is weak evidence of AI search capability.

Here is what sits under the hood when an LLM decides who to cite:

1. Crawling: AI engines access your pages through various crawlers and retrieval methods.
2. Chunking and embedding: the engine splits your page into passages and converts them to vector embeddings rather than scoring the document as a whole. Dense retrievers beat keyword matching by 9 to 19 points on passage retrieval tasks, per Karpukhin et al..
3. Retrieval: retrieval-augmented generation pulls the passages most semantically relevant to the query, as described by Lewis et al..
4. Grounding: models cross-check claims against independent sources and reward information consistency, per Google's AGREE research.
5. Citation: models cite passages that are extractable and corroborated, and ignore the rest.

So when you read reviews, you're really evaluating whether an agency understands web search, AI citations, and training data as connected surfaces. We cover the mechanics in same channel, two surfaces and in our video breakdown.

Comparing review source reliability#

Four sources dominate agency vetting, and they differ on three attributes that matter to you.

No single row wins on all three attributes, which is why weighting matters more than collection.

Filtering noise from actionable reviews#

Readable reviews contain operational specifics: named methodology, team composition, measurement cadence. A review that mentions a documented framework, such as our CITABLE framework, tells you the agency runs a repeatable process. A review that says "great content, nice people" tells you the account manager was pleasant. Both may be true, but only one predicts pipeline.

Evaluating agency credibility on Clutch#

Clutch is the strongest public directory for agency vetting, provided you read it correctly.

How Clutch verification impacts rankings#

Clutch verifies reviews through industry-specific forms, direct client interviews, and unique review links sent to past clients. Its Leaders Matrix maps providers by focus and proven delivery. We see this as useful baseline verification: you know a project occurred and the client was real. You do not know whether the work moved pipeline, and you cannot trust category position as a quality signal.

Identifying pay-to-play review signals#

We have observed that Clutch sells sponsorship that increases visibility on category pages, so the agencies near the top of a category page often paid for greater exposure. Within sponsorship tiers, ranking is determined by Clutch Rank, which reflects reviews and reputation rather than additional spend. Read sponsored placement as advertising spend and review text as delivery evidence. Our own approach is the opposite trade: Discovered Labs is an organic search agency for B2B SaaS, built AEO-native, and we publish pricing and research openly rather than buying directory placement.

Evaluating AEO agencies via Clutch#

Read the review text for technical nouns: structured data, schema markup, passage optimization, citation rate, AI visibility tracking. When the text only mentions "content" and "communication," you are probably reading about a traditional SEO engagement. You can see the structure we mean on our verified Clutch profile, where clients describe the work in their own words. We hold a 4.9 out of 5 across 4 verified reviews.

Filtering Clutch reviews for quality#

Apply four filters before a review influences your scorecard:

• ARR match (Annual Recurring Revenue): prioritize reviewers from companies in your ARR band (for our ideal customer profile, $2M to $50M).
• Vertical match: a B2B SaaS review tells you more than generic small-business praise.
• Recency: AI search has changed rapidly, so recent reviews typically carry more weight.
• Technical detail: specific deliverables beat sentiment every time.

Evaluating G2 profiles for AEO agency selection#

G2 is the weakest of the four sources for agency vetting, and the reason is structural.

The software bias in G2 reviews#

G2's taxonomy emphasizes software attributes like features, ease of use, integrations, and pricing tiers. Those prompts map poorly to professional services, where the product is a team and a process. As a result, agency profiles on G2 tend to be sparse, and the reviews that do exist often evaluate a tool the agency uses rather than the agency's work. Omnius markets a proprietary AEO attribution tool, which fits G2's software format neatly, while we focus on reporting in pipeline: MQLs, demos, and revenue-adjacent metrics.

G2 signals for agency vetting#

If a shortlisted agency does have a G2 profile, two data points are worth extracting:

• Quality of support: how responsive the vendor's team is when something breaks.
• Ease of doing business: per G2's research guidelines, this reflects how users rate working with the vendor overall, and we read it as a rough proxy for onboarding and contract friction.

Skip the aggregate star rating. Sample sizes are too small to mean anything for services.

Private peer feedback: high-signal, no records#

Why referrals carry higher pipeline signal#

Private referrals skip every marketing filter that corrupts public reviews. There is no star rating to game, no profile to sponsor, no marketing team approving the message. A peer CMO will tell you what broke, how the agency handled it, and whether the work actually showed up in pipeline. That honesty is what directories cannot sell you.

Assessing referral quality in agency vetting#

Ask peers five questions before you weight their referral:

1. Who was actually on the team: full-time AI/ML engineers, or SEO strategists reading from a playbook?
2. How did they handle attribution discrepancies between GA4, HubSpot, and the CRM?
3. What did monthly reporting look like: a narrative tied to pipeline, or a data dump?
4. Were contracts month-to-month, or did they push an annual lock?
5. What would you do differently if you hired them again?

Mapping peer referrals to MQLs#

Referral behavior is the strongest review signal an agency can earn, and occasionally it becomes public. Tom Wentworth, CMO at incident.io, made this public in our published case studies:

"I have recommended you to multiple peer CMOs. There are large organizations like Hubspot and Ramp who have dedicated teams to work on large projects like AEO. For everyone else (except my competitors) there's Discovered Labs!"

That statement exists because the work produced numbers worth referring: incident.io booked 22% more organic meetings and grew AI visibility from 38% to 64%, closing the gap on PagerDuty. When a review documents both the referral and the result, you have found your highest-signal input.

Filtering noise from unverified vendor feedback#

The marketing bias in user reviews#

Self-published testimonials are marketing assets: the agency chose them, edited the context, and published only the engagements that ended well. That doesn't make them useless. You audit them rather than read them.

Evaluating vendor-hosted case studies#

A high-signal case study contains three things: the specific technical approach, exact unrounded metrics, and a clear attribution path. Our Discovered Labs reviews page pairs each number with the method that produced it. Gladia grew sales-accepted leads 7x in 4 months, with 93% of AI-referred leads coming from LLM search. Sova Assessment lifted organic demo requests 167%, and organic became their #1 channel for leads and MQLs. One anonymized B2B SaaS client grew AI-referred trials 6x in 7 weeks to 3,500+ trials. Each engagement ran the same motion: content restructured for passage retrieval, plus off-page information consistency. You can watch the mechanics in our B2B SaaS case study video.

We expand that into a six-point audit in what a credible AEO case study contains.

Identifying high value attribution signals#

The strongest case studies show how pipeline was captured, not just that it grew. We tell buyers to look for self-reported attribution fields ("how did you hear about us"), UTM parameters on AI-referred sessions, and CRM syncing that ties those sessions to MQLs and opportunities. We detail the full measurement stack in our post on proving AI search ROI.

How to rank review sources by signal quality#

Assigning weights to review sources#

Build your scorecard around where honesty lives. In our approach, we prioritize:

• Peer referrals: Unfiltered, no commercial incentive.
• Verified case studies with attribution paths: Numbers you can audit.
• Clutch reviews: Verified but commercially influenced.
• G2 and unverified testimonials: Baseline color only.

Mapping review signals to pipeline#

A high-signal review proves the agency understands CAC, payback periods, and board-level reporting, not just rankings. When a reviewer mentions MQL-to-opportunity conversion or marketing-sourced pipeline, that agency has sat in meetings like yours. When they mention only traffic, they have not.

Identifying low quality review signals#

Discount any agency whose reviews or site lean on guaranteed rankings in 30 days, content velocity boasts, hidden pricing, or 12-month lock-ins. Transparency is a signal in itself, which is why we keep our public pricing on the site:

A Search Visibility Diagnostic runs €4,370 one-off, credited in full to your first retainer if you sign within 14 days.

Evaluating data sources for agency vetting#

Aggregating AI and SEO review signals#

Reviews should cover both surfaces of organic search: rankings and traffic on Google, plus citation rate and share of voice in AI answers. The two overlap less than most teams assume: 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. Our own analysis of 144,000 AI citations found Reddit referenced in roughly 27% of ChatGPT search results while appearing in 0.35% of visible citations. That gap is why our Reddit marketing services run as citation infrastructure, not a social play.

This is also where our proprietary AI visibility measurement earns its keep: it tracks citation rates and share of voice across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews, and clients get view-only access to the same dashboards we work from. The six audit layers behind it are documented in what a GEO (Generative Engine Optimization) audit covers.

Parsing Clutch ratings for signal#

Run this sequence on any agency's Clutch profile. First, ignore the overall rating entirely. Second, open the three most recent reviews and read only the Project Feedback and Results sections. Third, highlight every technical noun and every number. Fourth, check whether the reviewer's company matches your ARR band and vertical. The rating is marketing. The text is evidence.

Real-world AI pipeline data examples#

AI visibility measurement is probabilistic, and any agency that claims otherwise is overselling. We published a measurement flaw in tracking platforms showing why these tools estimate rather than count. In the anonymized engagement above, custom UTM parameters and CRM syncing tied AI-referred trials to pipeline, which is the level of proof your scorecard should demand.

Assessing credibility in AI SEO agency reviews#

Ranking agency reviews by reliability#

The hierarchy, from highest to lowest signal:

1. Private peer referrals with named results.
2. Vendor case studies with exact metrics and attribution paths.
3. Clutch verified reviews from ICP-matching companies.
4. G2 profiles, where they exist.
5. Unverified testimonials and directory badges.

Minimum review counts for agency vetting#

You don't need hundreds of reviews. In our view, a handful of detailed, verified reviews from companies matching your ideal customer profile (ICP) typically tell you more than a hundred generic five-star ratings, because review volume past that point usually reflects collection effort rather than delivery quality.

How to verify agency credibility#

Reading reviews correctly is step one. Verifying the numbers inside them is step two, and we set out that protocol in how to verify an AEO agency's claimed results.

Run this checklist before signing anything:

1. Verify claims across three surfaces: search the brand on Google, ask ChatGPT and Claude about them, and check whether third-party sources tell a consistent story.
2. Audit their own visibility: an agency selling AI search should be cited by AI search. Our AI search audit guide walks through the process, and what an AEO audit covers lists the deliverables.
3. Score their content: run their blog through our free AEO content evaluator and see whether they practice passage-ready structure.
4. Check entity clarity: their brand, founders, and methodology should be unambiguous to a machine (clear entity relationships in copy and schema markup).
5. Run the free baseline: our free AI visibility checklist gives you a citation-rate baseline before any sales call.

For your own site's LLM readability, we prioritize elements like answer-first openings on key pages, validated schema markup, sections that independently answer one question, consistent facts across your site and third-party profiles, and visible timestamps on everything you publish.

How to weight internal agency reviews#

Reviews of the work matter less than reviews of who did it. A credible AI search partner employs full-time AI/ML engineers alongside SEO strategists and editors. Our team is led by Ben Moore, an ex-Stanford AI researcher who built fraud systems at Stripe and Coinbase, and Liam Dunne, a growth operator who scaled a B2B SaaS past $20M ARR. We pair editorial-quality content standards with in-house AI engineering and proprietary visibility tracking, and we run SEO and AI search as one operation across web search, citations, and training data.

Build your shortlist on signal, not scores#

Public directories give you baseline verification, and that is all they give you. The decision itself should rest on peer referrals and case studies with attribution paths you can audit, the only sources that prove an agency connects organic search to pipeline. The partner you want runs SEO and AEO as one operation, reports in MQLs and demos, and lets you leave any month. Before you commit to anyone, request a Search Visibility Diagnostic: €4,370 one-off, credited in full to your first retainer if you sign within 14 days. It gives you an honest baseline of your brand's visibility across Google and AI engines, with no obligation. You can see the content methodology behind our work in our CITABLE framework post.

FAQs#

How much does a technical AEO retainer cost?#

Our Establish tier starts at €7,995 per month on a month-to-month basis, covering up to 20 content units, AI visibility tracking, and technical SEO. A 6-month commitment drops the rate to €6,995. Each content unit represents a piece of optimized content structured using our CITABLE framework.

How long does it take to see AI citation results?#

Initial citations typically appear within 1 to 2 weeks of publishing optimized content. A material lift in citation rate across priority queries usually takes 3 to 4 months.

Do you require long-term contract commitments?#

No. Every retainer is month-to-month with 30 days notice, and longer commitments exist only to lower your rate.

Are Clutch reviews reliable for evaluating AEO agencies?#

Clutch verifies that projects happened through client interviews, but category rankings reward paid profiles. Treat Clutch as baseline verification and weight peer referrals and verified pipeline data higher, as detailed in the review source weighting section above.

Is Discovered Labs associated with medical testing laboratories?#

No. Discovered Labs is an organic search agency for B2B SaaS, with no affiliation to clinical or medical diagnostic laboratories.

Key terms glossary#

AEO (Answer Engine Optimization): The practice of optimizing content to appear in AI-generated answers from engines like ChatGPT, Claude, Perplexity, and Google AI Overviews, distinct from but complementary to traditional SEO.

GEO (Generative Engine Optimization): Another term for optimizing content for AI-powered search and answer engines, focusing on how large language models retrieve and cite information.

ICP (Ideal Customer Profile): The description of a company or buyer persona that would derive the most value from your product or service, typically defined by attributes like ARR, industry vertical, and company size.

MQL (Marketing Qualified Lead): A prospect who has shown sufficient interest and fit to be passed from marketing to sales for follow-up.

ARR (Annual Recurring Revenue): The yearly value of recurring subscription revenue, a key metric for SaaS businesses.

Citation rate: The percentage of priority buyer queries where an AI engine explicitly cites your brand as a source. It is the AI-search equivalent of share of search.

Passage retrieval: The process by which an LLM extracts a specific, semantically relevant block of text from a page to answer a query. Pages built from self-contained, answer-first sections win more retrievals.

Information consistency: The alignment of accurate claims about your brand across independent third-party sources. LLMs use that agreement to verify facts before citing you.

CITABLE framework: Discovered Labs' 7-component content-for-retrieval framework covering entity clarity, intent architecture, third-party validation, answer grounding, block structure, freshness, and schema. It structures pages so LLMs can quote, verify, and cite them.