AEO 101Single source of truth on AEO

For ecommerce and DTC brands

Be the brand AI recommends when shoppers ask which product to buy.

When a shopper asks ChatGPT for the best in your category, AI names two or three brands. Either your SKU is one of them, or you are not in consideration.

By Subia Peerzada/Updated May 14, 2026

§01 How does the new buying funnel actually work?

A shopper types one prompt. AI fans out to a knowable source pool. Two or three brand names land in the answer.

01The buyer prompt
>_
ChatGPT, best running shoes for plantar fasciitis under $150
02Retrieval fanout
retaileramazon.com
reviewrunrepeat.com
reviewwirecutter.com
forumreddit.com/r/running
ownedyourbrand.com
03The named answer
Top picks for plantar fasciitis under $150:
Hoka Bondi 8 — heel cushion + arch support
Brooks Glycerin 21 — neutral max-stack
ASICS Gel-Kayano 30 — guidance pillar

The brand name on the highlighted line is not earned by the PDP. It is earned by what the cited sources say.

Your PDP rarely makes the cut on its own. The brands that get named are the ones whose SKU shows up correctly on the pages AI already trusts.

§02 What happened to the old buying funnel?

The compare-tabs middle of the shopping funnel is gone. AI does the compare for the buyer.

Pre-AI funnel2023
  1. Shopper Googles the product category
  2. Opens 4 to 8 product pages
  3. Reads a Wirecutter or Reddit thread
  4. Compares prices on Amazon and DTC
  5. Reads reviews
  6. Picks a SKU

6 steps

AI-answer funnel2026
  1. Shopper asks ChatGPT for the best product
  2. AI names two or three brands
  3. Shopper checks out

3 steps

The compare-and-research middle of the funnel collapsed into the AI answer. Either you are named in the answer, or the shopper never sees you.

Top-of-funnel and discovery now compress into a single moment. Win that moment or lose the cart.

§03 Which sources does AI actually read from?

AI shopping answers come from a knowable source pool. Influence over that pool is the actual work.

The source pool AI reads from

What we influence, tier by tier

01 · Tier 1Retailer aggregatorsamazon.com · walmart.com · target.com · sephora.comHighest citation weight on commercial product queries. Where your SKU lives, how its attributes are written, and how reviews land here moves more visibility than any PDP rewrite.
02 · Tier 1Independent review siteswirecutter.com · runrepeat.com · rtings.com · gear patrolEditorial third-party validation. Inclusion in a category roundup is one of the most reliable predictors of AI recommendation.
03 · Tier 2Forum and community threadsreddit.com · quora.com · category-specific subforumsAI fans out to these for first-person opinion and edge cases. Authentic, sustained presence wins. Pay-to-play does not.
04 · Tier 2Category specialists and trade presscategory trade pubs · vertical newsletters · category aggregatorsNarrative authority. The pubs your category trusts shape the framing AI uses when describing your space.
05 · Tier 3Owned PDPs and brand siteyourbrand.com/product/skuLower weight in shopping answers than most brands assume. Still required as the canonical SKU record AI cross-checks for attribute drift.

Tier 1 sources carry most of the weight on commercial queries. Tier 3 carries the record. We work all three, not just the brand site.

The shopper never sees this list. The model does, every time. Get inside it, on the right tier.

§04 What metric actually decides the category?

Citation share on one named prompt, broken out by surface. Every category has its working set.

Citation share visualisation

Prompt: best running shoes for plantar fasciitis under $150

Category defaultChallengerLong tail
ChatGPT48% · 22% · 30%
Hoka
Brooks
Claude42% · 27% · 31%
Hoka
Brooks
Perplexity38% · 24% · 38%
Hoka
ASICS
Gemini35% · 26% · 39%
Hoka
Brooks
AI Overviews41% · 23% · 36%
Hoka
Brooks

Illustrative shares from a representative prompt audit. Real engagements run 80 to 200 prompts weekly with named-platform deltas.

The job is to know your distribution, name the competitors, and move the bars on a weekly cadence.

§05 What do we actually ship?

Six lines of work, run weekly, owned by us.

Each block describes the actual work, not a tool we hand over. You receive named outcomes on a fixed cadence; we carry the production, the retailer relationships, and the platform monitoring.

01

SKU-level visibility audit on the prompts buyers actually ask

We map shopper prompts to your SKUs and audit how AI answers today. The output is a SKU-by-SKU read of where you appear, where a competitor wins, and where the answer pool has not yet chosen.

02

AI shopping prompt curation, owned and maintained

Your category has 80 to 200 prompts that decide the sale. We curate the list on day one and re-run it every week against ChatGPT, Claude, Gemini, Perplexity, and AI Overviews.

03

Product attribute reconciliation across the source pool

AI pulls product attributes from your PDPs, retailer pages, marketplace listings, and review sites. When those disagree, AI picks one. We reconcile the attributes that decide recommendation and keep them aligned across surfaces.

04

Retailer-page and marketplace citation work

ChatGPT cites Amazon, Walmart, Sephora, and category aggregators more than your DTC site. We brief retail teams on the structures that get pulled and place your SKUs inside the answer pool the way AI reads it.

05

AI product recommendation surfacing across categories

Recommendation rate is the metric that matters: not whether AI mentions you, but whether AI recommends you over a named competitor. We engineer the third-party validation that moves it.

06

Weekly drift monitoring and course correction

Catalogs change. Retailer prices change. Reviews change. AI answers drift with them. We monitor the curated prompt set weekly and respond inside seven days when a source-pool shift hits a SKU you care about.

§06 The methodology is public

One framework, applied weekly. Research, playbook, and engineering ledger all open.

The methodology

CITE framework

Comprehend, Influence, Track, Evolve

The living playbook

AEO 101

Refreshed daily from our research library

The research

AI Visibility Index

Open library of operator-grade briefs

§07 Questions buyers ask before they engage

The questions ecommerce buyers ask before they engage.

How does ChatGPT decide which products to recommend?
ChatGPT pulls from a small set of cited sources for any shopping query. The pool usually includes retailer pages (Amazon, Walmart, Sephora, category specialists), independent review sites, Reddit and forum threads in the category, and a handful of editorial publications. Your own PDPs are rarely the deciding source. The model weighs product attributes, review consistency, freshness, and the structure of the cited pages. Brands win recommendation when their products appear with consistent attributes and credible third-party validation inside that source pool, not when they spend more on the PDP itself.
Why don't my PDPs appear in AI shopping results?
Two common reasons. First, the PDP is built for Google ranking, not for passage extraction. AI pulls clean attribute blocks, concise feature passages, and structured FAQ content; most PDPs bury that under marketing copy. Second, AI rarely cites the brand's own product page when a higher-authority retailer or review site carries the same SKU. The fix is not to compete with the retailer page. The fix is to make sure your SKU is correctly represented on the pages AI does cite, and to engineer your own PDP so that when AI does read it, the answer-grade content is at the top.
Does Amazon or major retailer presence help or hurt?
It mostly helps, with one caveat. Retailer presence puts your SKU inside the pages AI already cites, so it raises the floor on visibility. The caveat is attribute drift. If your Amazon listing says one thing and your DTC PDP says another, AI picks the version it trusts more, which is usually the retailer. The work is making sure your SKU information across Amazon, Walmart, Sephora, Target, and the category specialists agrees with your own site on the attributes that decide recommendation: size, materials, ingredients, compatibility, warranty.
How do AI Overviews change product discovery?
AI Overviews now sit above the blue links on a large share of commercial queries. For product searches, the Overview names two to five products by brand, summarises attributes, and links to a handful of cited pages. If your SKU is not in that Overview, the click rate on the page-one organic result below it drops sharply, and the shopper may never scroll. The implication for ecommerce is that ranking position one on Google is no longer the same as being the first answer; the first answer is the Overview, and it has its own rules.
What can ecommerce do that traditional SEO cannot?
Traditional SEO optimises for ranking position, traffic, and click-through rate. AI visibility optimises for citation share, recommendation rate, and source-pool position on a curated set of shopping prompts. The work overlaps in places (schema, page structure, page authority) and diverges in others: retailer-page coordination, third-party review density, forum and Reddit positioning, comparison-page engineering, and attribute consistency across surfaces. A good SEO program will not produce these outcomes on its own. They need a separate operating discipline.

More vertical playbooks from Cite Solutions

PR & Comms→Consumer Apps→Professional Services→Automotive→Travel→Dev Tools→

Ready to become the answer AI gives?

Book a 30-minute discovery call. We'll show you what AI says about your brand today. No pitch. Just data.

Book a Discovery Call
.md