Skip to content
NewConfidence intervals on every metric

Know what AI says about you. Then change it.

Manifest Signal asks ChatGPT, Perplexity, Gemini, and Claude the questions your customers ask, records every answer and source, and turns what it finds into prioritized, evidence-linked work.

14-day trial · No card required · Built by Manifest FTS

Observes the answer engines your buyers already use

  • ChatGPT
  • Perplexity
  • Gemini
  • Claude
  • AI Overviews
  • Copilot

The shift

Buyers ask for answers now. Most teams can’t see them.

When someone asks an AI assistant which provider to choose, there is no results page to check and no rank to track. There is a paragraph that names a few organizations and cites a few sources, and it changes from one request to the next.

Answers, not rankings

Visibility is whether you are named, where, and how you are described, measured across many samples rather than a single position.

Sources decide

Engines lean on a small set of pages and third-party sites. Knowing which ones tells you where effort will count.

Errors repeat

An outdated price or wrong location can be repeated to every buyer who asks. You need to see it before they do.

How it works

From question to measurable change

Signal connects observation, diagnosis, and action in one loop, so every recommendation can be traced back to what engines actually said.

  1. 01

    Define the questions

    Track the prompts your buyers actually ask, grouped by topic and intent, alongside the competitors they compare you with.

  2. 02

    Observe every engine

    Signal asks ChatGPT, Perplexity, Gemini, and Claude on a schedule and stores each full answer, its sources, and the model that produced it.

  3. 03

    Diagnose with evidence

    See where you are missing, which sources shape answers, what engines get wrong, and whether crawlers can reach your pages.

  4. 04

    Act and measure

    Work a prioritized task list where every item links to its evidence, then watch the metrics that matter over the next runs.

Methodology

Numbers you can defend in front of leadership

AI answers vary. A dashboard that hides that variation invites bad decisions. Signal shows its work.

  • Every rate has a denominator

    Sample size sits next to every percentage, in the product and in reports.

  • Intervals, not guesses

    95% Wilson intervals; a change is flagged only when intervals separate.

  • Raw answers on record

    Open any number to read the exact answers and sources behind it.

  • Sample data is labeled

    Illustrative data never mixes with live observations.

Read the methodology

Free tool

Can AI crawlers actually read your site?

Check robots.txt rules for eleven AI crawlers, indexability, structured data, and server-rendered content in about ten seconds. No account needed.

Pricing

Simple plans that scale with your program

Every plan includes readiness audits, evidence-linked tasks, and the full methodology. Start with a 14-day Growth trial.

Starter

$49/month

25 prompts · 3 engines · weekly runs

Growth

Most popular

$149/month

100 prompts · 6 engines · daily runs

Agency

$399/month

300 prompts · 6 engines · daily runs

FAQ

Questions teams ask before they start

Can’t find an answer? Contact us or read the documentation.

What is Manifest Signal?

Manifest Signal is a web application from Manifest FTS that measures how AI answer engines such as ChatGPT, Perplexity, Gemini, and Claude describe and cite an organization. It asks each engine the questions your customers ask, records every answer and source, reports visibility with sample sizes and confidence intervals, checks your site’s technical readiness for AI crawlers, and turns gaps into prioritized, evidence-linked tasks.

How is this different from an SEO rank tracker?

Rank trackers record a position in a list of links. Answer engines return a written answer that may name several organizations, cite a handful of sources, and change from one request to the next. Signal stores the full answer, measures mention, citation, position, and accuracy across many samples, and reports rates with confidence intervals instead of a single rank.

Which answer engines do you track?

ChatGPT, Perplexity, Gemini, and Claude through their official APIs with web search enabled. Google AI Overviews and Microsoft Copilot are available as labeled sample data while official access is evaluated. Every observation records the model identifier.

Can you guarantee that AI engines will recommend us?

No, and you should be wary of anyone who does. Signal shows what engines say today, why, and which changes are most likely to help. It reports results honestly, including when a change is within normal variation.

Is the data in a new workspace real?

New workspaces start with sample data so every screen is useful immediately. Sample data is labeled everywhere it appears and is never mixed into live metrics. Switch to live collection in workspace settings when live engines are enabled.

Do you need access to our website or analytics?

No. Signal only needs your brand name, domain, competitors, and prompts. Readiness audits fetch public pages the way a crawler would.

How long is the trial, and do I need a card?

Fourteen days with Growth plan limits. No card is required. If you do not choose a plan, the workspace becomes read-only; nothing is deleted.

Who builds Signal?

Signal is built and supported by Manifest FTS, an engineering-led digital partner that designs, builds, and maintains web platforms. Teams that want hands-on help can engage Manifest FTS for content, technical, and structured-data work.

Find out what AI says about you this week.

Start with sample data in minutes, invite your team, and switch to live observations when you are ready.

14-day trial · No card required · Cancel anytime