Due Diligence Guide Updated April 2026

AEO Agency Red Flags: 11 Warning Signs for Ecommerce Brands 2026

The AEO market is full of SEO agencies who rebranded a deck and started charging more. These are the 11 things that expose them fast. Read this before your next agency call.

11 Warning Signs
6 Critical Flags
3 High Flags
2 Medium Flags
01

The 11 Red Flags

01
They cannot explain variant canonical handling
The single most reliable filter for ecommerce AEO competence
Critical

Ask them: when a product has 12 colour and size variants each with their own URL, what is the canonical strategy and how does that affect LLM retrieval? A real AEO agency for ecommerce answers this immediately with specifics about self-referencing canonicals, parameter handling, consolidation logic, and how retrieval pipelines treat the relationship between parent and variant pages. An agency that does not do this work will pause, then pivot to something vague about "making sure Google understands the content". That pivot is the flag.

What the bad answer sounds like
"We ensure all your product pages are properly optimised for AI search."
"We handle canonical tags as part of our technical audit."
Changing the subject to page speed or meta descriptions.
What a real answer includes
Self-referencing canonical on each variant with consolidation to parent for retrieval.
How variant URLs affect ProductGroup vs Product schema structuring.
What happens when Googlebot and a retrieval crawler treat the same URL differently.
Ask in the pitch call: "Walk me through your canonical strategy for a Shopify product with 15 variants. What does that do to LLM retrieval of the parent product page?"
02
No citation tracking tools or methodology
If they cannot measure it, they are not doing it
Critical

AEO without citation tracking is just content production with a new name. Ask to see a live dashboard or report showing how a current client's brand appears in ChatGPT, Perplexity, and Google AI Overviews responses for target queries. Ask how frequently it is updated and what methodology they use to track prompt variations. Agencies doing real work have built or bought tooling for this. Agencies faking it will send you a screenshot from a single ChatGPT prompt they ran the morning before the call.

What the bad answer sounds like
Showing one ChatGPT screenshot as a "citation report".
"We monitor AI mentions manually on a monthly basis."
GA4 organic traffic as the proxy for AEO performance.
What a real answer includes
Named tooling: Profound, Semrush AI Toolkit, custom prompt monitoring, or proprietary builds.
Prompt variation methodology covering query type, phrasing, and user intent segments.
Separate tracking across ChatGPT, Perplexity, and Google AIO with different cadences.
Ask in the pitch call: "Can you show me a citation tracking report from an active ecommerce client? What tool produces it?"
03
Their AEO deliverable is blog content
Content programs are one component. Agencies leading with them are selling the wrong thing
Critical

Ask them to break down the first 90 days of an ecommerce AEO engagement. If the answer is primarily a content calendar, a blog rollout, or a publishing schedule, you are talking to a content agency. For ecommerce, the first 90 days of real AEO work is schema audit and remediation, variant URL architecture review, canonical health, structured data implementation on product and category pages, and retrieval testing. Blog content supports AEO. It is not AEO. Agencies that pitch the content plan as the strategy have not built the technical foundation the content would sit on.

What the bad answer sounds like
"Month one: keyword research and content brief creation."
"We'll publish 8 AEO-optimised blog posts per month."
A 90-day plan with no schema or technical deliverables.
What a real answer includes
Schema audit with specific findings on Product, Offer, and AggregateRating markup.
Retrieval baseline: how does the brand currently appear across LLM platforms?
Technical remediation before any content production begins.
Ask in the pitch call: "What are the first three non-content deliverables in your ecommerce AEO onboarding?"
04
No ecommerce product schema case studies
Portfolio proof matters more in AEO than almost any other discipline
Critical

Ask for a documented case study where schema implementation work on a product catalogue drove a measurable citation or visibility lift. If they cannot produce one, they have not done the work at scale. Testimonials do not count. Slides summarising outputs do not count. A real case study identifies the baseline schema state, what was changed, what structured data types were implemented or corrected, and what the citation or traffic outcome was over a defined window. The absence of this is not a minor gap — it means they are learning on your budget.

What the bad answer sounds like
"Our clients have seen great results, we can share a testimonial."
A deck slide with percentage uplifts and no methodology attached.
"Schema is part of our technical hygiene process."
What a real answer includes
Named client or anonymised industry, with before and after schema state documented.
Specific schema types implemented: ProductGroup, Offer, AggregateRating, etc.
Measured citation frequency or organic visibility change over a defined period.
Ask in the pitch call: "Can you walk me through a specific schema implementation you did for an ecommerce client and what the measured outcome was?"
05
They cannot distinguish how ChatGPT, Perplexity, and Google AIO retrieve differently
Treating all LLMs as one target is a foundational strategy error
Critical

Each major AI platform retrieves, weights, and cites differently. Perplexity runs live web retrieval and is faster to reflect schema changes. ChatGPT draws on training data plus Bing-indexed content with browsing enabled. Google AI Overviews is deeply entangled with existing Search signals and E-E-A-T. A strategy that treats all three as one surface is not a strategy — it is a guess. If your agency cannot explain the material differences and how their work accounts for them, they built their AEO practice on surface reading of a few blog posts.

What the bad answer sounds like
"We optimise for all AI search platforms."
No distinction made between platforms in their reporting or strategy docs.
"AI SEO works the same way regardless of the tool."
What a real answer includes
Explicit per-platform retrieval logic and why it matters for product content.
Different measurement cadences per platform based on update frequency.
Platform-specific content structuring decisions explained with rationale.
Ask in the pitch call: "How does your strategy for Perplexity retrieval differ from your Google AI Overviews strategy for a product category page?"
06
FAQ schema is their primary AEO recommendation
FAQ markup is entry-level. Agencies leading with it have not gone further
Critical

FAQ schema is real, useful, and completely insufficient as an AEO strategy for ecommerce. It addresses one surface — question and answer extraction — and ignores the product data architecture that drives the majority of ecommerce LLM citations. Brands get cited for product information: price accuracy, availability, specifications, review aggregation, comparison-ready attributes. None of that is FAQ territory. Agencies that lead with FAQ schema in their AEO pitch are showing you the edge of their knowledge, not a strategy.

What the bad answer sounds like
"We add FAQPage schema to product pages to get into AI answers."
FAQ sections as the main structured data deliverable.
No mention of Product, Offer, or AggregateRating markup anywhere in the proposal.
What a real answer includes
Product and Offer schema treated as the core retrieval infrastructure.
AggregateRating and Review markup as trust and comparison signals.
FAQ schema positioned as a supplementary signal, not the strategy.
Ask in the pitch call: "Outside of FAQ schema, what structured data types are you implementing on product and category pages and why?"
07
No process for inventory signal freshness
Out-of-date availability and pricing data actively suppresses retrieval
High

LLMs and retrieval crawlers penalise or deprioritise product pages where the structured data for price, availability, and offer validity is stale or inconsistent with live page content. For ecommerce brands with dynamic catalogues, this is a constant maintenance problem. Ask the agency how they handle Offer schema freshness across a catalogue of 5,000 SKUs. Ask whether they have a monitoring process for availability status discrepancies between what the schema says and what the page renders. If they have no answer, they are not thinking about ecommerce at the product data layer.

What the bad answer sounds like
"The client's dev team handles inventory data."
No awareness that schema data freshness affects retrieval signals.
"We set the schema once and monitor it in Google Search Console."
What a real answer includes
Monitoring cadence for price and availability schema vs live page state.
Integration or handoff process with the client's product data pipeline.
Understanding of how stale Offer data affects AI Overview eligibility.
Ask in the pitch call: "How do you handle keeping Offer schema data current on a catalogue with frequent price and availability changes?"
08
No real Shopify or headless stack experience
Generic schema advice breaks on the platforms most ecommerce brands actually use
High

Shopify has specific constraints around product schema that generic advice ignores: liquid template rendering, the way metafields interact with structured data, variant URL generation, and the limitations of Shopify's default schema output. Headless stacks introduce additional complexity around SSR, hydration, and how schema gets injected at build versus runtime. Agencies that cannot speak to these specifics are working from textbook knowledge, not deployed experience. For Shopify Plus brands in particular, the implementation details are not optional — they are where the work lives.

What the bad answer sounds like
"We can implement schema on any platform."
No mention of Liquid, metafields, or platform-specific rendering behaviour.
Treating Shopify and a custom React build identically in the proposal.
What a real answer includes
Specific Shopify schema limitations and how they route around them.
Metafield strategy for surfacing custom product attributes in structured data.
Headless rendering approach: where schema is injected and how it is validated.
Ask in the pitch call: "What are the specific schema limitations in Shopify's default output and how do you address them for AEO?"
09
Reporting shows organic traffic, not citation metrics
What gets measured reveals what actually gets done
High

Ask to see a sample monthly report from an ecommerce AEO client. If the dashboard is predominantly organic sessions, keyword rankings, and impressions from Google Search Console, you are looking at an SEO report with a new title. Real AEO reporting tracks citation frequency across LLM platforms, brand mention position within AI responses, structured data validation scores, competitor citation comparison, and which query categories are driving retrieval. Organic traffic is a downstream output. Agencies that report on it as the primary AEO metric are not tracking the actual work.

What the bad answer sounds like
A GA4 and Search Console dashboard rebranded as an "AI visibility report".
Keyword ranking tables as the core deliverable.
"We track AI search impact through organic traffic trends."
What a real answer includes
Citation frequency by platform, query category, and competitor comparison.
Schema validation metrics and structured data error tracking.
Named AI Overviews appearance data, not just inferred from traffic patterns.
Ask in the pitch call: "Can I see a sample report? Show me where the citation tracking data lives in it."
10
No competitor citation analysis in the audit
AEO without a competitive retrieval baseline is optimising blind
Medium

Before any ecommerce AEO strategy makes sense, you need to know which competitors are currently getting cited for your target product queries and why. Schema structure, content format, domain authority signals, and entity associations all contribute. Agencies that jump straight from onboarding to execution without producing a competitor retrieval landscape are skipping the diagnostic that shapes the strategy. You end up with a generic playbook instead of one built around closing the specific gap between your citations and your competitors.

What the bad answer sounds like
Onboarding audit covers technical health, not competitor citation positions.
"We'll work on improving your AI visibility across key terms."
No mention of who is currently winning retrieval for the client's core queries.
What a real answer includes
Audit output showing which brands get cited for the client's target product queries and on which platforms.
Analysis of what the cited competitors are doing technically that the client is not.
Gap analysis feeding directly into the prioritised implementation roadmap.
Ask in the pitch call: "What does your competitor citation audit look like and when does it happen in the engagement?"
11
They promise citation results within a few weeks
Realistic timelines separate honest agencies from those closing deals
Medium

Perplexity and ChatGPT with browsing can reflect new content relatively quickly, sometimes within weeks for fresh pages. But for ecommerce product schema remediation, category page architecture changes, and the compound effects of retrieval optimisation at catalogue scale, the honest measurement window is 90 to 180 days. Google AI Overviews typically takes longer. Agencies that promise visible citation lifts in two to four weeks are either measuring something cosmetic, cherry-picking a single favourable query, or setting you up to see early noise and renew before the real measurement window closes. Honest agencies give realistic timelines and explain why.

What the bad answer sounds like
"You'll start seeing AI search results within 2 to 4 weeks."
30-day success metrics built into the contract.
No mention of seasonality, platform update cycles, or measurement methodology.
What a real answer includes
Per-platform timelines: Perplexity fastest, Google AIO slowest, ChatGPT in between.
Distinction between early retrieval signals and meaningful compound effects.
Seasonality accounted for in the measurement window and baseline methodology.
Ask in the pitch call: "What is a realistic timeline to see measurable citation movement on Google AI Overviews for a product category page, and how do you account for seasonality in that measurement?"
02

All 11 Flags at a Glance

# Red Flag Severity What It Reveals The Test Question
01 Cannot explain variant canonical handling Critical No real product schema depth Explain canonical strategy for 15-variant product
02 No citation tracking tools Critical Not measuring what they claim to do Show me a live citation report
03 AEO deliverable is blog content Critical Content agency, not AEO specialist Name 3 non-content 90-day deliverables
04 No ecommerce schema case studies Critical Unproven on product data work Show a schema case study with measured results
05 Cannot distinguish ChatGPT vs Perplexity vs AIO Critical Surface-level AEO knowledge only How does Perplexity strategy differ from Google AIO?
06 FAQ schema is the main AEO recommendation Critical Entry-level knowledge, no product layer What structured data beyond FAQ are you implementing?
07 No inventory signal freshness process High Not thinking at product data layer How do you handle Offer schema freshness at scale?
08 No real Shopify or headless experience High Generic advice, platform-specific failure What are Shopify's schema limitations and your workarounds?
09 Reporting shows rankings, not citation metrics High Measuring SEO, not AEO Show me where citation data lives in your report
10 No competitor citation analysis in audit Medium Generic playbook, no competitive baseline What does your competitor citation audit include?
11 Promises citation results in weeks Medium Closing deals, not setting realistic expectations What is a realistic Google AIO timeline and why?
03

Frequently Asked Questions

The most critical flags are inability to explain variant canonical handling, no citation tracking tools or methodology, pitching blog content as AEO, zero ecommerce product schema case studies, and treating FAQ schema as the main AEO deliverable. Any agency that cannot walk you through how a product variant gets handled in an LLM retrieval pipeline is selling SEO rebranded as AEO.

Ask them to explain the difference between how ChatGPT, Perplexity, and Google AI Overviews retrieve and rank product content. Ask what happens to a product variant URL in their schema architecture. Ask them to show you a citation tracking dashboard from an active ecommerce client. Ask for a documented case study where AEO work caused a measurable citation lift on a specific product category. Vague answers to any of these mean you are talking to an SEO agency calling itself AEO.

No. FAQ schema is one minor signal in a complete AEO architecture. For ecommerce, the structural work is in Product, Offer, AggregateRating, and ProductGroup schema, variant-level canonical handling, inventory signal freshness, and category page retrieval logic. Agencies that lead with FAQ schema as their AEO play are either early in their learning curve or are repackaging basic on-page SEO work. It is a starting point, not a strategy.

04
The Agency That Clears All 11

Northquery passes every test on this list

All 11 flags above come from real gaps we found when evaluating agencies for ecommerce AEO. Northquery scored 88.8 out of 100 across seven technical criteria and clears every flag here with documented work to back it up.

Variant canonical strategy documented with Shopify Plus case work
Citation tracking built into every engagement from day one
Technical first 90 days with schema audit before any content begins
Ecommerce schema cases with before and after measurement published
Per-platform strategy with separate cadence for each LLM target
Honest timelines with seasonality-adjusted measurement windows
See the full Northquery methodology
Full scoring methodology with founder disclosure, case studies, and reasoning published at northquery.com.