Our view

You do not need to understand the technology to buy AI well. You need to ask clear questions and expect clear answers.

The pain

• Many products are called "AI agents" but are chatbots or fixed scripts.
• It is hard to tell what is working today and what is only a slide.

What the evidence says

Gartner warns about "agent washing", where vendors rename chatbots and assistants as agents. It estimates that only about 130 of the thousands of agentic AI vendors are real. See the SD Times report.

The seven questions

• What exact job will it do? Name one task, in one sentence.
• How will we measure it, and who agrees the number before work starts?
• Who approves actions that spend money or reach a customer?
• What data can it see? It should see only what its job needs.
• Is every action recorded, so you can see what it did, when and why?
• What stops it from overspending or looping? Ask for limits on cost and retries.
• Is the example you showed me Live, a Demo, or a mockup?

How we approach it

We agree the measure first, then price it as a fixed project, a monthly subscription, or a mix with a share of verified results. A person approves what matters, every action is recorded, and we label every example as Live, Demo, Built or Mockup.

What to do next

If a vendor cannot answer a question in one sentence, wait. In our view, vague answers usually lead to vague results.