New: the 2026 Holiday AI Recommendation Share Index. See what AI recommends →
Client Login Discover Your Score
Knowledge Hub · Research

Research & Indexes

BrandRank's AI Recommendation Share Indexes are syndicated studies of what AI engines recommend. Each index puts the questions real customers ask to the leading AI engines, scores every answer, and shows which retailers, brands and products AI names first, and why.

Open the latest index How the indexes work

Latest index

950AI answers scored
170holiday questions
6AI engines

Engines tested: ChatGPT, Claude, Gemini, Grok, Perplexity, Meta

LatestSyndicated Index

AI's Most Recommended Holiday Gifts & Retailers

The BrandRank 2026 Holiday AI Recommendation Share Index℠

Six AI engines answered 170 holiday shopping questions: 160 single-answer questions and 10 open ones. They largely agree on where to shop but not on what to give: on the single-answer questions, four or more engines named the same answer on 65% of retailer questions and just 31% of product questions.

Explore the interactive report →

Key findings

  • Three retailers own the answer. Amazon (125), Target (78) and Nordstrom (72) take 29% of all 950 scored answers, and each owns a different moment: Amazon the screen, Target the store and pickup, Nordstrom service and returns.
  • A few gifts win; most are wide open. Only one product question is unanimous (Amazon Fire HD Kids). Brands fare better: Uniqlo, Nike, Levi's, Bellroy, Quince, Patagonia and LEGO each win a unanimous verdict.
  • The engines disagree, sharply. Gemini never names Walmart first. Perplexity names Nordstrom 20 times and Amazon 11; Gemini names Amazon 26 times.
  • Customer service is judged by returns. Nordstrom takes 25 of 48 core service answers; Amazon takes one. No engine credits anyone for AI-powered support.
  • Agents and accuracy are the next test. Four engines name Amazon the best store for an AI agent to buy from, and four still call its assistant Rufus, a name retired in May 2026.
Method

How every AI Recommendation Share Index works

Each index follows the same four steps, so results can be compared across engines and over time. Each report publishes its own question count, engines and dates.

  1. Ask

    The same questions real customers ask go to every engine under an identical protocol, in the same window of time.

  2. Answer

    For each single-answer question, every engine names exactly one retailer, brand or product, with a short rationale. Engines may decline to answer.

  3. Normalize

    Names are grouped by product family for consensus and by parent company for Answer Share, so "Kindle" and "Ring" count toward Amazon.

  4. Check

    BrandRank flags answers that are outdated, inaccurate or off-question, and reports where the engines agree and where they split.

Answer Share
The share of scored answers that name a given retailer, brand or product. It counts first-choice recommendations, not mentions.
Consensus
A question has consensus when four or more of the six engines name the same answer. Unanimous means all six agree.
The Answer Economy
The shift from searching to asking, where AI answers increasingly decide which brands people choose. The term comes from Pete Blackshaw's book of the same name.
Your brand, your category

See what AI says about your brand

Custom Answer Reports use the same method on your own category, products and competitors, including how AI describes your customer service.

Request an Answer Report Email us
Questions

About BrandRank research

What is an AI Recommendation Share Index?

An AI Recommendation Share Index is a syndicated BrandRank.AI study of what AI answer engines recommend in a category or season. BrandRank puts the same set of real customer questions to several leading AI engines under an identical protocol, scores every answer, and reports which retailers, brands and products each engine names first, where the engines agree and where they split.

Which AI engines does BrandRank research cover?

Each report lists the engines it tested. The 2026 Holiday AI Recommendation Share Index covered six: ChatGPT, Claude, Gemini, Grok, Perplexity and Meta.

Are the results product tests or rankings of real-world performance?

No. The indexes measure what AI engines say at the time of the study. They are not product tests and do not measure how retailers or brands actually perform. They are intended for brand intelligence, not shopping advice.

How do I cite BrandRank research?

Cite the index name, BrandRank.AI and the month of the study, and link to the report page. For example: BrandRank.AI, The BrandRank 2026 Holiday AI Recommendation Share Index, October 2026, brandrank.ai/2026-holiday-ai-recommendation-share-index.

Can BrandRank run this research for my brand or category?

Yes. BrandRank builds custom Answer Reports around a brand's own category, products, competitors and customer service reputation. Request one or email contact@brandrank.ai.

Stay current

Get the next index first

One email when we publish new research. No digest spam, unsubscribe anytime.

Looking for articles, news and press releases? Browse Prompted Perspectives & News →