AI answer audit · SalesBuilder

What you get

20 questions · Twenty buyer questions in: who ChatGPT, Perplexity, Gemini and Google’s AI Mode name and cite for each, and whether any page of yours answers it. · Keyword
share · Answers that name you, over all answers: 0.04 is 4 of every 100. A new domain reads 0 until a cited page of its own exists.
questions[].answers · For each engine: whether it names you, the brands it names, and the pages it cites.
questions[].yourPage · Your page that answers the question, with the quote that shows it does, or answers: false.
readiness · Whether robots.txt lets answer engines read the site, whether llms.txt exists, and the schema each key page carries.
citedDomains · The domains answers cite most, with counts: the pages to be listed on, or to pitch.

How long: About 2 minutes; ChatGPT answers take about 30 seconds each, asked in parallel.

Trigger

It runs when you call it. Search readout re-asks the saved questions every week.

TypeScript

every.week(($) => $.Keyword.track({ site: 'leaseline.app', terms: 'saved', questions: 'saved' }))

Steps

Your agent does these steps, one row per step; each step’s check is part of its row and its price.

Step	Verb	Returns	Price	Approval
1. Ask each engine each question three times	`Keyword.ask`	Each answer: the brands it names and the pages it cites	About 96¢	Runs on its own
2. Read the 20 pages answers cite most	`Page.audit`	Each cited page’s content, for what it says that yours does not	About 0.4¢	Runs on its own
3. Audit your key pages for answer readiness	`Page.audit`	robots.txt, `llms.txt`, schema and the first-60-words answer test on 10 pages	About 0.2¢	Runs on its own
4. Check whether your page answers each question	`Domain.verify`	`answers: true` with the quote code finds on your page, or `answers: false`	About 1¢	Runs on its own
5. Count share and the gaps	`Collection.query`	Share across every answer, and the questions no page of yours answers	Free	Runs on its own

Calls: Keyword.ask, then Page.audit, then Domain.verify, then Collection.query.

About 98¢ per 20 questions, each time it runs, capped at $1.50.

Example: give it a goal

Check whether ChatGPT, Perplexity, Gemini and Google’s AI Mode name Leaseline for the 20 questions landlords ask most, spend up to $2, and show me the plan first.

The plan it shows runs 5 steps, priced before it spends anything; Run it above shows the result it returns.

When something fails

Each failure, and what happens next

When	Then
An engine refuses or times out on every route	It is left out of `share` and named under `skipped`; share is never computed over a missing engine.
An engine’s three answers disagree	All three are kept on the Event, and share counts each; in goal mode the agent may ask that engine twice more and records the Decision with its cost.
Your site blocks answer engines in robots.txt	`readiness.robotsAllowsAnswerEngines` reads `false` and the readout leads with it: no page fix matters until crawlers can read the site.
You call it again the same day	It returns the stored audit at $0, keyed on questions, site, engines and day.

Each step retries three times, then falls through to the next route; one item's failure lands in skipped with its reason, never failing the call. Before a step would cross maxCost it stops and returns what it has, with a resume link. Each step's idempotency key is the call's key plus the step and the item: leaseline-keyword-audit-1-<step>-<item>.

Change it

Ask	Compiles to
Add Claude and Copilot to the engines.	`Keyword.audit`
Write an answer page for every question no page of mine answers.	`Page.answer`
Ask these questions again every week and tell me what changed.	`Keyword.track`

TypeScript

import { $, workflow } from 'startup-builder'

export default workflow({
  name: 'keyword-audit',
  verb: 'Keyword.audit',
  input: { questions: 'string[]', site: 'string', engines: 'string[]', runs: 'integer' },
  maxCost: 1.5,
  execute: async (ctx, { questions, site, engines = ['chatgpt', 'perplexity', 'gemini', 'google-ai-mode'], runs = 3 }) => {
    const answers = await ctx.step('answers', () => $.Keyword.ask({ questions, venues: engines, runs, maxCost: 1.1 }))
    const cited = await ctx.step('cited', () => $.Page.audit({ urls: topCited(answers, 20), fields: ['content'], maxCost: 0.02 }))
    const ready = await ctx.step('readiness', () => $.Page.audit({ site, pages: 10, checks: ['robots', 'llms', 'schema', 'first-answer'], maxCost: 0.01 }))
    const yours = await ctx.step('yours', () => $.Domain.verify({ hostname: site, answers: questions, fields: ['quote'], maxCost: 0.02 }))   // a checked extraction
    return auditOf(questions, answers, cited, ready, yours)
  },
})

Extend it with extends: 'Keyword.audit' and base(), or compose your own from these steps.

Works with

Plays that order it: Answer engine optimization, Make your API easy for agents to find, Developer relations, Free tools.

It serves Check AI answers, and runs in Launch and search. It deepens Content marketing.

For agents

For agents: the guide, its approvals, retries and lookups

Fact	Value
Guide	`Keyword.audit`
Guide approval	`limits`
Goal	Check whether ChatGPT, Perplexity, Gemini and Google’s AI Mode name Leaseline for the 20 questions landlords ask most, spend up to $2, and show me the plan first. It may change counts, the order of steps, soft filters and the route a read takes, and it stops early when the goal is met. It never relaxes your words, raises its own budget, answers its own approval, calls a Play or acts outside your standing approval.
Trigger	`every.week(($) => $.Keyword.track({ site: 'leaseline.app', terms: 'saved', questions: 'saved' }))`
Retry	Three times with backoff, then the next route; the same idempotency key never charges, sends or posts twice
Idempotency key	`leaseline-keyword-audit-1`
Lookups	A `find:` input asks unless exactly one record fits

Step
1. Ask each engine each question three times
2. Read the 20 pages answers cite most
3. Audit your key pages for answer readiness
4. Check whether your page answers each question
5. Count share and the gaps

The amounts live once, on the hub (approvals).

Who does each step, and how it is checked

Step	Function	Doer	Checker	Assurance	Inputs
1. Ask each engine each question three times	`code`	Code	Code checks each result against its declared type	`measured`	`$0.questions, $0.engines, $0.runs`
2. Read the 20 pages answers cite most	`code`	Code	Code checks each result against its declared type	`measured`	`$1.cited`
3. Audit your key pages for answer readiness	`code`	Code	Code checks each result against its declared type	`measured`	`$0.site`
4. Check whether your page answers each question	`agentic`	The Content lead agent	Code finds the quoted evidence on the page	`measured`	`$0.site, $0.questions`
5. Count share and the gaps	`code`	Code	Code checks each result against its declared type	`measured`	`$1, $2, $3, $4`

The doer is never the checker, and the weakest assurance on the path is the whole call’s.

Under the hood

Under the hood: the Systems, Nouns and Events it touches

Kind	What it touches
Systems	None.
Nouns	Keyword, Page, Domain, Collection
Events	`Domain.verified`, `Collection.queried`

Reference

Inputs, outputs, its Event and its errors

Name	Type	Default	Meaning
`questions`	`string[] \| { fromMap: KeywordMap; take: number }`	Required	The questions buyers ask, in their words, or the question-form terms from a Keyword opportunity map.
`site`	`Hostname`	Required	Your site.
`engines`	`VenueId[]`	`['chatgpt', 'perplexity', 'gemini', 'google-ai-mode']`	The answer engines asked. Claude adds about 2.5¢ a question.
`runs`	`integer`	`3`	Asks per question per engine. Answers vary between asks; share is read across all of them.
`maxCost`	`number`	`1.50`	The most this call spends each time it runs, in USD. It stops before a step would cross it and returns what it has, with a `resume` link.

Emits Keyword.audited. Refuses with max-cost before it spends, and with the Data calls and Accounts groups (refusals).

What you getCopy Anchor Link

TriggerCopy Anchor Link

StepsCopy Anchor Link

Example: give it a goalCopy Anchor Link

When something failsCopy Anchor Link

Change itCopy Anchor Link

Works withCopy Anchor Link

For agentsCopy Anchor Link

Under the hoodCopy Anchor Link

ReferenceCopy Anchor Link

Contents