See how clear information becomes a useful answer.

Edit the example information and ask a question. This preview uses X-Protocols data.

The example business

Nordvik Bilservice is invented for this preview. These are the information items it publishes.

Item 1
Item 2
Item 3
Edit as a table
First line is the column names. Tab or comma separated, up to 20 items. Rename a column and the fields above rename with it.

The question

Ask it the way a customer would ask an assistant.

Answeredresolved

One published information item answers that question.

Answered from

  • serviceTyre swap
  • locationMunich
  • price40 EUR
  • notePer wheel, seasonal

What the assistant is toldState these values exactly. Do not round them, combine them, or extrapolate beyond them.

Preview only. This answer is not signed, not stored and not published.

View technical response
{
  "status": "resolved",
  "reason": "One published information item answers that question.",
  "agent_instruction": "State these values exactly. Do not round them, combine them, or extrapolate beyond them.",
  "columns": [
    "service",
    "location",
    "price",
    "note"
  ],
  "matches": [
    {
      "location": "Munich",
      "note": "Per wheel, seasonal",
      "price": "40 EUR",
      "service": "Tyre swap"
    }
  ]
}

Try this with my business

What this preview leaves out

A published profile adds the parts a preview cannot have: a freshness date a customer can read, updates from the systems you already use, a signed response an assistant can check, and pages AI assistants read directly. This preview stores nothing, signs nothing and publishes nothing.

How this works once it is published