What is prompt tracking?

Prompt tracking is running a fixed set of buyer questions against AI assistants on a schedule and reading what comes back. A prompt is one question; a topic is a named group of prompts. Attensira runs each tracked prompt against each tracked model, stores the answer verbatim, and reports two things from it: whether the answer named your brand, and whether it attached a link to your own domain.

How does prompt tracking work?

You define the prompts that matter to your business. Attensira runs each prompt against each tracked model, at most once per prompt, model and country per day, and stores the answer text. It then reads each stored answer for a mention of your brand and a citation of your domain, buckets every cited domain by kind, and computes the same rates for the competitors you name. Today's numbers are always partial, because the day's runs are still being collected.

How many prompts can I track?

Prompt slots are a stock, not a monthly allowance: nothing resets, and removing a prompt frees its slot immediately. The slot count for each of the three plans is on the pricing page — there is no fourth plan and no custom tier. Start with 10-20 questions your buyers would actually type, not keyword variants.

Can I track competitors for each prompt?

Yes. You can add any competitor domain to your workspace. For every prompt you track, Attensira shows a side-by-side comparison of your mention rate versus each competitor. Competitor rates share your denominator, so they do not sum to 100% — several brands can be named in one answer, or none.

How is this different from traditional keyword tracking?

A keyword tool reports a position in a list. There is no position here: an answer either names you or it does not, and either attaches a link to your domain or does not. So the readings are two rates over the same denominator — mention rate and citation rate — each carried with the number of successful runs behind it, and neither reported as changed until a two-proportion z-test clears 95%.

Can I export my prompt tracking data?

Yes. Every table and chart the agent draws carries Copy, which puts the rows on your clipboard as TSV, and Download, which saves the same rows as CSV. The same numbers are also readable over the MCP server and the REST API, so you can pull them into your own stack instead of exporting by hand.

Prompt Tracking

Know What AI Tells Your Customers

AI doesn't rank you 1 through 10. You're either in the answer or you're invisible. Track your mention rate, your citation rate, and competitive position across ChatGPT, Claude, Google AI — updated daily.

Get Started Book A Demo

•Cancel anytime•Setup takes under 2 minutes

Track Your Presence

See if AI mentions you or your competitor when customers ask about your category.

Compare with competitors

The same mention rate for every competitor you name, on the same denominator as yours.

Uncover Citations

See exactly which websites AI cites when it recommends you — or someone else.

Benchmark Daily

Five model surfaces, run at most once per prompt, model and country per day. Today’s numbers are always partial.

app.attensira.com

Pipedrive

Overview

Mention rate

61.2%

+5.2% vs last period

Citation rate

28.4%

+3.1% vs last period

Total Mentions

2,847

+412 vs last period

Visibility Trend

Pipedrive

Industry Avg

Mon

Tue

Wed

Thu

Fri

Sat

Sun

Visibility by Model

ChatGPT

68%

Claude

62%

Perplexity

57%

Google AI

51%

Readable model surfaces; a new workspace starts on two

1/day

Max runs per prompt, model and country

24h

Refresh cycle

n=1/n=3

Draws per reading, by plan

Why This Matters

Keywords had Google. Prompts have AI.

For two decades, SEO meant tracking how your domain ranked for keywords on Google. You had a position — #1, #5, #20 — and you optimized to move up. That model is breaking down.

When someone asks ChatGPT “what's the best CRM for startups?” there's no list of 10 blue links. AI synthesizes a single answer, mentioning some brands and ignoring others entirely. Your customers are already asking AI before they search Google — and you have no idea what it tells them.

Prompt tracking is how you measure your presence in this new landscape. Just like you tracked keyword rankings daily, you now need to track how AI models respond to the prompts your customers actually ask — across every tracked model and country.

The shift from keywords to prompts

Traditional SEO“best CRM” → Rank #4

AI Prompt Tracking“best CRM for startups” → Mentioned

AI doesn't rank you 1-10.

You're either in the answer — or you're not.

How prompt tracking works

Define the prompts your customers ask, and we'll show you exactly how every AI model responds — every day.

Add Your Prompts

Enter the questions your customers ask AI — like "best CRM for startups" or "Salesforce alternatives for small teams." These are your tracked prompts, the AI equivalent of your target keywords.

Start with 10-20 high-intent prompts for the strongest signal

We Query AI Models Daily

Attensira runs each tracked prompt against ChatGPT, Perplexity, Claude, Gemini and Google AI — five surfaces, no more — and stores the answer verbatim so the evidence stays readable. Google AI has no API reader, so expect thinner data there.

Five surfaces, at most one run per prompt, model and country per day

Measure and Compare

See the mention rate, the citation rate, the sources cited and the run count behind each, for every prompt. Add a competitor by domain and the same two rates are computed over the answers already collected.

Mention rate, citation rate and sources — each with the number of runs behind it

Everything you need to track AI visibility

Monitor how AI models talk about your brand across every major platform and country.

61.2%

mention rate · n=25

Pipedrive

+5.2 pts vs. last week · clears 95%

Mention rate

Successful runs that named your brand, over all successful runs in the window, with the run count attached. No change is reported until it clears a 95% significance test.

Pipedrive

61%

Salesforce

58%

HubSpot

45%

Competitor rates

The same mention rate for every competitor you name, on the same denominator as yours. They are independent rates, not slices of a pie, so they can sum past 100%.

g2.com/pipedrive

capterra.com/crm

pipedrive.com

Citation Sources

See which websites AI models cite when mentioning your brand. Find the sources that drive your visibility — and the ones you're missing.

ChatGPT

68%

Claude

62%

Google AI

57%

Google AI

51%

By Model

Mention and citation rate are reported per model surface, never blended. Five surfaces are readable — ChatGPT, Perplexity, Claude, Gemini and Google AI — and Google AI has no API reader, so expect thinner data there.

🇺🇸United States

64%

🇩🇪Germany

58%

🇬🇧United Kingdom

52%

By Region

A prompt is run per country, so the same question can be answered differently in each. Rates are reported per country rather than averaged into one number.

Mention rate18% · n=21

Citation rate9% · n=21

Week-on-week deltano proven change

Rates, with the sample behind them

Every rate carries its run count, and a week-on-week delta is only called a change once it clears a two-proportion z-test at 95%. Otherwise it reads “no proven change” — which is not the same as zero.

What You Measure

What you measure for every prompt

For every prompt you track, Attensira captures the full picture: which brands get named, how often, whether a link to your domain came with it, and which sources were cited. The answer text is stored verbatim, so every finding can be re-checked.

Trend lines, competitor side-by-sides, and the sources cited — broken down by model and country, never blended into one score. Nothing here scores how warmly an answer spoke about you.

Mention rate

Successful runs that named your brand over all successful runs in the window, with the run count beside it.

Citation rate

Successful runs where a model attached a link to your own domain, on the same denominator as mention rate.

Competitor rates

The same two rates for the domains you name, scored over answers already collected. No extra runs, no prompt slots.

Citation Mapping

Which websites AI references when it mentions your brand. Build authority on the sources that matter most.

FAQ

Frequently Asked Questions

Have more questions?

Book a demo and we'll answer them personally.

Book a Demo

Prompt tracking is the AI equivalent of keyword tracking in traditional SEO. Instead of monitoring your Google rankings for keywords, you monitor how AI models like ChatGPT, Claude, and Google AI respond to specific prompts. You see how often your brand is named, how often your own domain is linked, and the same two rates for the competitors you name.

Five surfaces: ChatGPT, Perplexity, Claude, Gemini and Google AI. A new workspace starts with ChatGPT and Perplexity tracked, and you add the rest. Gemini is the one surface Starter does not include; it unlocks on Growth. Google AI has no API reader at all, so expect thinner and less regular data there than on the API-backed channels.

You define the prompts that matter to your business — questions your buyers ask AI. Attensira runs each prompt against each tracked model, at most once per prompt, model and country per day, and stores the answer verbatim. It then reads each answer for two things: a mention of your brand and a citation of your domain, and computes the same two rates for the competitors you name. Today’s numbers are always partial, because the day’s runs are still being collected.

For every prompt you get a mention rate (the share of successful runs where a model named your brand), a citation rate (the share where it linked your own domain), the sources behind those answers with each domain’s share of the citations, and the same two rates for the competitors you name. Every rate carries the number of runs behind it, and a prompt we never measured reads "not measured" rather than zero. There is no blended visibility score and no sentiment metric.

It depends on your plan. Our Starter plan includes 50 prompts, Growth includes 150, and Business supports 350 prompts. Enterprise plans offer custom limits. We recommend starting with 10-20 high-intent prompts that match how your customers ask AI about your category.

Yes. You can add any competitor domain to your workspace, and Attensira scores it over the answers it has already collected — no extra runs and no prompt slots consumed. Competitor rates share your denominator, so they do not sum to 100%: one answer can name four brands, or none.

Traditional SEO tools track your Google ranking position for keywords. Prompt tracking measures whether AI models mention your brand at all — and what they say about you. AI doesn't return a ranked list of 10 links; it gives a single synthesized answer. Your goal shifts from ranking #1 to being part of that answer.

Yes. Every table and chart the agent draws carries Copy, which puts the rows on your clipboard as TSV, and Download, which saves the same rows as CSV. The same numbers are readable over the MCP server and the REST API, so you can pull them into your own stack instead of exporting by hand.

Tracking is just the start

Once you know where you stand, use the rest of the Attensira platform to improve your AI visibility and build the presence that gets your brand recommended.

Content Optimization

Attensira does not score a page. It reads the pages behind a finding and proposes the specific edits, which reach you as a pull request.

The agent

It reads your tracked prompts, models, answers and citations, works out what changed, and acts on it — scoped to one workspace and stopped for an explicit yes before anything irreversible.

GitHub

A first-party GitHub App, scoped to the repositories you select. Every fix arrives as a pull request, and Attensira never merges it.

What it does not do

A mention is a case-insensitive substring match of your brand name or domain against the answer text. There is no entity resolution and no word-boundary check, so a brand called Arc matches “search”. Read a few stored answers before quoting a rate.
Today is always partial. Each Prompt × Model × country runs at most once per workspace-local calendar day, so compare whole days against whole days.
No visibility score and no sentiment metric exist in Attensira. Nothing here scores whether an answer spoke about you kindly.
It cannot attribute. A mention is not a visit and a citation is not a signup — there is no identity join between an answer a model gave someone and a person who later arrived at your site.

The full account is at docs.attensira.com/reference/limitations.

Which plans include it

Included on every plan: Starter, Growth and Business.

What varies by plan is sampling depth — one draw per reading on Starter, three on Growth and Business — plus how many Prompts you can track and which model surfaces are available. Those numbers live on pricing, and the behaviour is documented at docs.attensira.com/tracking/prompts.

By Karl-Gustav KallasmaaLast updated 4 September 2026

See what AI tells your customers

Start with the prompts that matter to your business. Each one runs at most once per model and country per day, so the first whole day of collection is your first comparable reading.

Get Started Book A Demo

Cancel anytime

Know What AI Tells Your Customers

Track Your Presence

Compare with competitors

Uncover Citations

Benchmark Daily

Overview

Keywords had Google. Prompts have AI.

How prompt tracking works

Add Your Prompts

We Query AI Models Daily

Measure and Compare

Everything you need to track AI visibility

Mention rate

Competitor rates

Citation Sources

By Model

By Region

Rates, with the sample behind them

What you measure for every prompt

Mention rate

Citation rate

Competitor rates

Citation Mapping

Frequently Asked Questions

What is prompt tracking?

What AI models do you track?

How does prompt tracking work?

What metrics can I track for each prompt?

How many prompts can I track?

Can I track competitors for each prompt?

How is this different from traditional keyword tracking?

Can I export my prompt tracking data?

Tracking is just the start

Content Optimization

The agent

GitHub

What it does not do

Which plans include it

See what AI tells your customers

Overview

Keywords had Google. Prompts have AI.

How prompt tracking works

Add Your Prompts

We Query AI Models Daily

Measure and Compare

Everything you need to track AI visibility

Mention rate

Competitor rates

Citation Sources

By Model

By Region

Rates, with the sample behind them

What you measure for every prompt

Mention rate

Citation rate

Competitor rates

Citation Mapping

Frequently Asked Questions

What is prompt tracking?

What AI models do you track?

How does prompt tracking work?

What metrics can I track for each prompt?

How many prompts can I track?

Can I track competitors for each prompt?

How is this different from traditional keyword tracking?

Can I export my prompt tracking data?

Tracking is just the start

Content Optimization

The agent

GitHub

What it does not do

Which plans include it

See what AI tells your customers