Can an AI agent actually shop these stores?
We scored 17 of the biggest Shopify brands on whether an AI shopping agent can read their products, check price and stock, and complete a purchase. Only 5 cleared 90.
60 seconds · free · any Shopify store
Key findings
What the scores reveal.
Elite ready
Only 5 stores cleared 90/100. The rest look agent-friendly; most aren't.
Upset
Size doesn't predict readiness. A small single-product brand out-scored billion-dollar names.
Universal blind spot
Stores ace inventory and cart, but write product pages for humans, not agents. The agent can't act on vibes.
Hidden risk
Even household names sit in the 60s. Policies and login often live in JavaScript an agent can't read.
Key findings (citable)
Copy-paste, with the source attached.
One-sentence facts from the June 2026 Daeri Agent-Readiness Leaderboard. Click any card to copy the stat with attribution.
The ranking
Agent-Readiness Score, 0–100.
The vertical line marks the group average. Bars to the right of it beat the field.
- 🥇AllbirdsReady98
- 🥈BeardbrandReady93
- 🥉Death Wish CoffeeReady92
- 4KnixReady92
- 5Rothy'sReady91
- 6Magic SpoonReady87
- 7SKIMSReady86
- 8GlossierReady86
- 9ColourPopReady85
- 10Kylie CosmeticsClose84
- 11Fashion NovaClose78
- 12BrooklinenClose78
- 13BombasClose77
- 14TentreeClose75
- 15GymsharkAt risk65
- 16MejuriAt risk65
- 17RuggableAt risk45
The full data
Five dimensions, one score.
Darker = stronger. Hover any row to scan it. Tap a dimension below for what it means.
| # | Store | Score | Discover | Product | Inventory | Action | Trust |
|---|---|---|---|---|---|---|---|
| 1 | Allbirds | 98 | 100 | 90 | 100 | 100 | 100 |
| 2 | Beardbrand | 93 | 98 | 82 | 100 | 100 | 83 |
| 3 | Death Wish Coffee | 92 | 100 | 70 | 100 | 100 | 100 |
| 4 | Knix | 92 | 100 | 66 | 100 | 100 | 100 |
| 5 | Rothy's | 91 | 100 | 65 | 100 | 100 | 100 |
| 6 | Magic Spoon | 87 | 100 | 57 | 100 | 100 | 83 |
| 7 | SKIMS | 86 | 75 | 60 | 100 | 100 | 100 |
| 8 | Glossier | 86 | 100 | 54 | 100 | 100 | 83 |
| 9 | ColourPop | 85 | 100 | 61 | 100 | 100 | 67 |
| 10 | Kylie Cosmetics | 84 | 100 | 47 | 100 | 100 | 83 |
| 11 | Fashion Nova | 78 | 100 | 65 | 100 | 78 | 50 |
| 12 | Brooklinen | 78 | 100 | 44 | 100 | 67 | 100 |
| 13 | Bombas | 77 | 75 | 55 | 100 | 67 | 100 |
| 14 | Tentree | 75 | 75 | 50 | 100 | 67 | 100 |
| 15 | Gymshark | 65 | 75 | 70 | 100 | 67 | 0 |
| 16 | Mejuri | 65 | 75 | 35 | n/a | 67 | 100 |
| 17 | Ruggable | 45 | 75 | 5 | n/a | 33 | 100 |
Average ~81. Inventory and cart are nearly universal; the points leak on product legibility and trust signals.
Why product legibility is the silent killer
An agent can't act on vibes.
It needs machine-usable facts: dimensions, materials, sizing — in structured data, not buried in marketing prose. Gymshark scored 0 on trust not because it has no return policy, but because its policies and login are JavaScript-rendered where an agent reading the page can't see them.
How we scored it
The same free audit you can run on your store.
Daeri runs a static, platform-agnostic audit the way an agent reads a store: llms.txt and sitemap, whether robots.txt blocks AI crawlers, product structured-data completeness (plus an LLM judgment of how concrete the copy is), variant-level price and stock, a machine-detectable cart, and reachable policies and login. Five weighted dimensions into one score. Every score here came from the same free public tool you can run on your own store.
A score is the floor
The real test is proof.
A score tells you your data looks ready; the harder question is whether an agent can actually complete the checkout. Daeri runs that too: a real browser agent attempts each task, multiple times, and reports the honest success rate.
Find your own number.
It takes 60 seconds, it's free, any Shopify store — and you can watch a real agent try to shop yours.