Prompt tracking
Rank tracking for the questions people ask AI.
Pick the prompts that matter to your business and search.co runs them across ChatGPT, Perplexity, Gemini, Claude and Google AI Overviews on your schedule. You see whether you were named, where in the list, which page was cited and how the answer changed since last time.
Ask Brigham.aiTrack “best pest control in Bentonville” daily in ChatGPT, Perplexity and Gemini and alert me if BentonvillePest.co drops out of the top three.| Prompt | ChatGPT | Perplexity | Gemini | AI Overviews |
|---|---|---|---|---|
| best pest control bentonville | #2 | #1 | #3 | Cited |
| termite inspection cost nw arkansas | Not named | Cited | Not named | Cited |
| eco friendly exterminator | #4 | #3 | Not named | Not named |
| mosquito treatment for yards | Not named | Not named | Not named | Not named |
- 5
- Assistants per prompt
- Daily
- Or weekly runs
- Or Slack alerts
- Full
- Answer history kept
Why track prompts
A prompt is the new keyword. Treat it like one.
- Assistants
- ChatGPT, Perplexity, Gemini, Claude, AI Overviews
- Cadence
- Daily or weekly per prompt
- Alerts
- Email or Slack
- Export
- CSV of every run
For twenty years, SEO teams have tracked keywords: pick the searches that bring buyers, check your position every day, act when it moves. AI assistants need the same discipline, but the unit is a prompt. “Who is the best pest control company in Bentonville?” is a question with a short list of answers, and the list changes.
Prompt tracking gives each of those questions a history. You see your list position per assistant, day by day, with the full text of every answer stored next to it. When a competitor appears above you or a citation disappears, you get an alert and the reason.
List position
Know where you sit in the answer, not just if you’re in it.
Most AI answers to buying questions are lists. The first name gets most of the attention, so moving from fourth to first is a real result even if your mention rate stays the same.
Each prompt gets a position chart per assistant, just like a keyword in a rank tracker. Competitors appear on the same chart, so you can see who you passed and who passed you.
What you get
Everything a keyword tracker does, built for prompts.
Scheduled runs
Run each prompt daily or weekly.
All five assistants
One prompt, checked in ChatGPT, Perplexity, Gemini, Claude and Google AI Overviews in the same run.
List position
Where you appear in the recommended list, tracked over time with competitors on the same chart.
Citation capture
The sources each answer links to, so you can see which of your pages are being used.
Alerts with reasons
Email or Slack when you enter or leave an answer, change position or lose a citation.
Answer history
Every answer stored in full, so you can compare today’s wording with last month’s.
Answer history
A diary of what each assistant said about you.
Open any prompt to see its timeline. Each run records the answer, the brands named, your position and the sources cited, with the changes since the previous run highlighted.
01
Diffs, not walls of text
New brands, dropped brands and changed sources are highlighted so you can scan a run in seconds.
At a glance
A dashboard for your prompt set.
Above the prompt list, four numbers summarize every tracked question: how often you are named, your average list position, how often you are cited and how many prompts moved since the last run.
Mention rate
46%
↑ +8 pts
Avg. position
2.6
↑ +0.9
Citation rate
21%
↑ +4 pts
Prompts lost
3
↓ since last run
How it works
Set up once, then let it run.
You can start from suggested prompts or bring your own list. Most projects are tracking in a few minutes.
- 01
Add prompts
Paste questions, import a CSV or send them over from prompt research in one click.
- 02
Group them
Organize by product, service area or funnel stage so reports roll up cleanly.
- 03
Pick assistants and cadence
Choose which assistants each group runs in and how often.
- 04
Set alerts
Decide which moves deserve a message: entering, leaving, position changes or lost citations.
Alerts you can act on
When something moves, Brigham tells you why.
An alert that says “you dropped” isn't much help. Each prompt alert links to a short Brigham summary: what changed in the answer, which sources were added or removed and what you can do about it.
Gemini’s latest answer cites two new pages with price tables. Your termite page has no pricing, and the reviews page it used to cite now redirects.
- 1Add a price range table to the termite page
- 2Restore or redirect the old reviews URL properly
- 3Re-run the prompt after the update
Ask Brigham
Ask about any prompt, in plain English.
Brigham reads the stored answers and their sources across all your tracked prompts and returns ranked fixes with the data behind each.
Six prompts name you in second or third place in at least two assistants. Four of them cite a competitor’s pricing or FAQ page that answers the question more directly than yours. These are your best short-term wins.
Recommended fixes
Sorted by impact
- 1
Answer the question in the first paragraph
30 minFor “eco friendly exterminator”, lead with the products you use and why, in two sentences.
- 2
Add an FAQ block to the service page
1 hrCited competitor pages answer cost, timing and safety questions. Yours answers one of three.
- 3
Create a mosquito treatment page
Half dayYou aren't named in any assistant for this prompt because you have no page on the topic.
Definitions
What each prompt metric means.
- Named
- The answer mentions your brand by name anywhere in the text.
- List position
- Your place among the brands the answer recommends. Blank if you are not named.
- Cited
- The answer links to one of your pages as a source, whether or not it names you.
- Movement
- Change in position or presence compared with the previous run of the same prompt and assistant.
Getting started
A good first prompt set.
Twenty to forty prompts is enough to see real patterns. Cover these kinds of questions.
Buying questions
4 checks- “Best [service] in [city]”
- “Top [product] for [use case]”
- “[Your brand] vs [competitor]”
- “Is [your brand] any good?”
Research questions
3 checks- “How much does [service] cost?”
- “How do I choose a [provider]?”
- “What should I ask a [provider]?”
Problem questions
3 checks- “How do I fix [problem]?”
- “Why does [problem] keep happening?”
- “When should I call a professional?”
How many prompts can I track?
As many as your credits allow. Each run of a prompt in an assistant uses a few credits, so many teams track their top prompts daily and the rest weekly within the 1,500 monthly credits on the $20 plan.
Do you run prompts from a specific location?
Yes. Local questions can be run for the location you set, which matters for “near me” and city prompts where answers change by area.
Where do the prompts come from?
You can write your own, import a list, or send prompts over from prompt research, which finds the questions people ask AI in your space grouped by intent.
Can I see the full answer, not just the score?
Yes. Every run stores the complete answer and its cited sources. You can read it, compare it with earlier runs and export it.
Start tracking the questions that sell.
Add your first prompts during the 7-day trial and see where you stand in each assistant after the first run.
Suggest 25 prompts my buyers likely ask AI assistants, start tracking them weekly in all five assistants, and alert me on big moves.