The Agent Shortlist Test

The Agent Shortlist Test — manual self-assessment
Use fresh chats in ChatGPT, Claude, Gemini and Perplexity; logged out where possible. Review what they actually say, then score yourself and two competitors against the author's rubric.
This is a manual self-check, not an automated scan, an industry benchmark, or a guarantee of recommendation. Results vary by model, account, location and time.
Your company details and answers stay in this browser unless you choose to copy prompts to an external assistant or use your email app to share results.

The ten prompts:
01 · Findable: What are the best {cat} vendors for {use} in {mkt}?
02 · Findable: Shortlist three {cat} providers for {use}, and explain why you chose each.
03 · Findable: Which {cat} tools would you recommend instead of {comp}?
04 · Readable: What does {co} do, who is it for, and what does it integrate with?
05 · Readable: Is {co} a good fit for {use}? What are its strengths and weaknesses?
06 · Provable: What results have customers achieved with {co}? Give specific numbers and customer names.
07 · Provable: Is {co} secure and compliant enough for a regulated company in {mkt}? Which certifications does it hold?
08 · Priceable: How much does {co} cost? Give a price range and the pricing unit.
09 · Priceable: Which is better value for {use}: {co} or {comp}? Compare the pricing.
10 · Overall: I have a budget for {cat} this quarter. Which vendor would you pick for {use} in {mkt}, and why?

The four gates and scoring rubric:
1. Findable — Can an agent discover you for the use case?
   0: Not named in any answer
   1: Named by some assistants, or only when asked directly
   2: Named in most category answers without prompting
2. Readable — Can it understand what you do, for whom, with what?
   0: Description wrong or missing
   1: Roughly right, but vague or out of date
   2: Accurate on product, customer and integrations
3. Provable — Can it verify that you deliver?
   0: No customers, numbers or certifications cited
   1: Customers or certifications named, without numbers
   2: Named customers with numbers, plus trust credentials
4. Priceable — Can it estimate what you cost?
   0: "Contact sales" or no idea
   1: Rough or outdated range
   2: Correct range and pricing unit

Author-rubric bands (only interpret a company's total after all four gates are scored):
7–8 · Shortlist-ready: Agents can find, understand, verify and price you.
5–6 · At risk: You appear, but one gate weakens the recommendation.
3–4 · Partly visible: Agents know you exist but cannot make the case for you.
0–2 · Invisible: You are unlikely to reach a human sales conversation through AI.