Public research, datasets and frameworks for understanding how AI systems discover, evaluate, select and recommend businesses.
Our mission is simple.
Understand how AI makes commercial decisions before everyone else.
- 22 public research studies
- 40,000+ AI recommendations analyzed
- 1,490 brands measured
- 5 commerce categories
- 66,000+ ecommerce stores analyzed
- 100% real-world data
- No surveys
- No simulations
- No synthetic datasets
Every report published here is based on captured AI responses, real ecommerce stores and controlled, reproducible experiments.
Our primary research synthesizing what we have learned about AI recommendation systems across commerce.
| Report | Description |
|---|---|
| The State of AI Recommendations Across Commerce 2026 | Cross-category analysis of 20,000 AI recommendations across five commerce industries. |
Controlled experiments designed to understand the mechanisms behind AI recommendation and selection.
- Web Search Rewrites 77% of AI Product Recommendations
- The Fame Study, Corrected
- AI Knows Your Website. It Still Won't Recommend You.
- 29,633 Reasons. 26,812 Unique. The Model Confabulates.
- Search Changes the Vocabulary, Not Just the Brands
- Candidacy vs Selection
- Nothing About Your Brand Predicts Recommendation. The Model's Own Past Behavior Does.
- Two Months Later, the Model Still Agrees With Itself
- Hand It a Rating, and It Follows Every Single Time
- We Invented a Brand With Zero History. Reviews Got It Picked Anyway
- We Widened the Fame Signal Four Ways. It Barely Moved
- The Model Hedges Most When It's Most Sure
These studies are designed to isolate individual variables and test whether observed patterns survive controlled changes.
Repeated recommendation experiments performed independently across different commerce categories.
- Beauty
- Supplements
- Coffee
- Pets
- Home & Living
Each category report follows the same core methodology, allowing results to be compared across industries.
Founder Lab is our public laboratory.
Instead of studying ecommerce brands only, we also build and document our own AI-native brand in public.
Current publications:
- Founder Lab — Day Zero
- Founder Lab — Research Log
The Founder Lab records changes, rescans and observations as they happen.
Atom Foundry researches how AI systems:
- discover businesses
- understand products
- build candidate sets
- select between alternatives
- generate recommendations
- explain their decisions
- respond to web search
- change their vocabulary and information retrieval
- maintain or change recommendation behavior over time
- respond to controlled changes in information and signals
The goal is not simply to measure whether a brand is visible.
The goal is to understand the mechanisms that determine which brands AI systems actually choose.
Every report follows the same principles.
- Real AI responses
- Real ecommerce stores
- Controlled experiments
- Public methodology
- Reproducible analysis
- No paid placements
- No sponsored conclusions
We separate observed behavior from interpretation.
If we cannot measure it, we do not publish it as a finding.
Our research combines several independent datasets and experimental approaches.
Including:
- AI recommendation experiments
- AI Commerce Score™ measurements
- Store architecture analysis
- AI readability evaluation
- AI understanding analysis
- AI trust analysis
- Cross-model comparison
- Controlled web search experiments
- Controlled signal manipulation
- Longitudinal rescans
- Candidate set and selection analysis
The research is built around recorded observations and controlled comparisons rather than assumptions about how AI systems work.
This repository supports the development of the AI Commerce Intelligence™ Framework.
Core concepts include:
- AI Readability™
- AI Understanding™
- AI Trust™
- Recommendation Intelligence™
- Recommendation Share™
- Recommendation Confidence™
- AI Commerce Score™
More information:
https://atomfoundry.dev/framework
Research begins.
First ecommerce stores scanned.
AI Commerce Intelligence™ introduced.
Cross-category recommendation experiments launched.
20,000 AI recommendations collected.
Initial mechanism studies published.
Controlled mechanism research expanded.
Longitudinal, search, selection and signal experiments added.
Research library expanded with additional mechanism studies and Founder Lab observations.
Research continues across recommendation intelligence, AI visibility, agentic commerce and AI decision science.
reports/
flagship/
mechanisms/
categories/
founder-lab/
framework/
methodology/
datasets/
observations/
timeline/
Website
Main website:
https://atomfoundry.dev
Research Library:
https://atomfoundry.dev/research
AI Commerce Intelligence Framework:
https://atomfoundry.dev/framework
Founder Lab:
https://founder.atomfoundry.dev
Citation
If you reference this work in research, articles, presentations or other publications, please cite the original report together with the publication date.
Example:
Atom Foundry. "The State of AI Recommendations Across Commerce 2026."
Atom Foundry Research, 2026.
https://atomfoundry.dev/research
For individual mechanism studies, please cite the original study title and its publication date.
License
This repository is released under the MIT License.
See LICENSE for details.