Independent AI product teardowns

Is it a

Productor just a Prompt?

We test AI apps against ChatGPT, Claude and Gemini—then expose the product layer that actually earns its place.

Scroll to evidence
LLM replaceability · Workflow depth · Domain UX · Data advantage · Action capability · Switching cost · Feedback loop ·
01

Fresh from the editorial desk

Latest teardowns

One framework. Seven dimensions. No automatic roasting.

Explore all 21 products
02

The verdict scale

Not every wrapper is weak.

We judge the product system around the model—not whether it uses one.

01

Just a Prompt

A strong prompt can replace nearly all of the core value.

02

Useful Wrapper

Helpful UX and workflow, but little durable advantage yet.

03

Real Product

AI is one component inside a meaningful product system.

04

Defensible AI Product

Workflow, data and feedback loops compound with usage.

03

Our method

Can it remember?
Can it act?
Does it improve?

We score the layers a chat window cannot naturally own: persistent workflow, domain interface, proprietary data, real-world actions, switching cost and feedback loops.

Read the framework

Put your product on the table

Think you built more
than a prompt?

Request a teardown