# A demonstration is not a working solution.

A tool can look convincing on a prepared example. A rollout decision needs ordinary inputs, difficult cases, constraints and the people who will use its output to be considered.

[Quality is difficult to judge](https://apiperson.com/ai/ai-evaluation/)

The answers look good, but acceptable quality has not been defined.

[The test needs real working data](https://apiperson.com/ai/ai-data/)

Information is scattered and access and permitted use remain unclear.

[Supplier proposals are hard to compare](https://apiperson.com/ai/ai-supplier-test/)

Each supplier gives a strong demonstration using its own examples.

All stories are fictional. They do not describe specific people, companies or projects. Any resemblance is coincidental.