Key takeaways
- AI citation tracking is not the same as rank tracking -- you need URL-level attribution to know which specific pages are being cited by ChatGPT, Perplexity, Claude, and other models
- The six metrics that matter most are: citation rate, brand visibility score, AI share of voice, mention-to-citation conversion rate, citation frequency by platform, and AI traffic attribution
- A well-built dashboard combines a dedicated AI visibility tool with a reporting layer (Looker Studio, Airtable, or similar) and a structured weekly review routine
- Most monitoring tools show you data but stop there -- the gap between "we were cited" and "here's the specific page losing to a competitor URL" is where most teams get stuck
- A 90-minute weekly review routine, run consistently, compounds over time and turns citation data into content decisions
If you've been paying attention to how people search in 2026, you already know that a significant chunk of your potential customers never see your website in a traditional search result. They ask ChatGPT. They use Perplexity. They get an answer from Google's AI Mode. And somewhere in that answer, either your brand gets cited or a competitor does.
The problem most teams run into: they know they need to track this, but they don't know what to build, what to measure, or what to actually do with the data once they have it. This guide walks through all three.
Why "we were mentioned in AI" is not enough
There's a version of AI citation tracking that feels productive but isn't. You set up a tool, it tells you your brand appeared in 34% of responses for a set of prompts this week, and you put that number in a slide deck. That's monitoring. It's not optimization.
The question that actually drives decisions is more specific: which URL on your site is getting cited for which prompt, and which competitor URL is beating it? That's attribution granularity. Without it, you can't prioritize content work, you can't measure whether a new article improved your citation rate, and you can't explain to a stakeholder why your AI visibility score went up or down.
So before you build anything, get clear on what you're trying to answer:
- Which of my pages are AI models citing, and for what prompts?
- Which prompts are my competitors visible for that I'm not?
- Is my AI visibility translating into actual website traffic?
- Which AI models are citing me most, and which ones are ignoring me?
Your dashboard exists to answer those questions, not to produce a number that looks good in a report.
The six metrics worth tracking
A lot of dashboards track everything and surface nothing useful. Here are the metrics that actually connect to decisions.
Citation rate is the percentage of AI responses (for a defined prompt set) that include a citation to your domain. It's your baseline visibility number. Track it per AI model -- your citation rate on Perplexity might be very different from your rate on ChatGPT, and they require different fixes.
Brand visibility score is a composite metric that weights citation rate by prompt volume. Being cited for a high-volume prompt is worth more than being cited for an obscure one. Most dedicated AI visibility tools calculate this automatically.
AI share of voice compares your citation rate against competitors for the same prompt set. You might have a 40% citation rate, which sounds fine -- until you see that your main competitor has 70% for the same prompts.
Mention-to-citation conversion rate is the ratio of brand mentions (where your brand is referenced in an AI response) to actual citations (where a specific URL from your site is linked). A high mention rate with a low citation rate means AI models know you exist but aren't pointing users to your content. That's a content structure and authority problem.
Citation frequency by platform breaks down which AI models are citing you and how often. Some brands are well-represented in Perplexity but invisible in Google AI Overviews. The fix for each is different.
AI traffic attribution connects citations to actual sessions on your site. This is the hardest metric to get right, but it's the one that justifies the budget. You need either a tracking code snippet, a Google Search Console integration, or server log analysis to see AI-referred traffic.
Choosing your core tracking tool
The tool you pick for data collection shapes everything else. Here's how the main options break down for citation tracking specifically.
Dedicated AI visibility platforms
These are purpose-built for tracking how AI models respond to prompts about your brand and category. They handle prompt querying, citation extraction, and competitor comparison.
Promptwatch sits at the more comprehensive end of this category. It tracks 10 AI models (ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews, Grok, DeepSeek, Copilot, Meta AI, Mistral), gives you URL-level citation data, and -- importantly -- includes an answer gap analysis that shows which prompts competitors rank for that you don't. It also has crawler logs showing which AI bots are actually visiting your site, which is genuinely useful for diagnosing indexing gaps.
For teams that want something lighter, a few other tools are worth knowing:
Otterly.AI is affordable and covers the basics -- brand mention tracking across major LLMs, share of voice comparisons, and prompt monitoring. It's a solid starting point if you're just getting into this space and don't need deep content optimization features.
Peec AI is strong on multi-language tracking, which matters if you operate in multiple markets. It covers the main AI platforms and gives you visibility scores per prompt.
Profound has a clean interface and good competitor comparison features. It's more monitoring-focused than optimization-focused, but the data quality is solid.
AthenaHQ covers 8+ AI search engines and is particularly useful for teams that need structured prompt management -- you can organize prompts by topic, funnel stage, or persona.
Tools with AI visibility as a secondary feature
Semrush has added AI visibility tracking to its platform, though it uses fixed prompt sets rather than custom ones. If you're already a Semrush user, it's a reasonable starting point before committing to a dedicated tool.
SE Ranking has an AI visibility module that integrates with its broader SEO toolkit. Useful if you want to track traditional rankings and AI citations in the same platform.
Comparison: what each tool covers
| Tool | URL-level citations | Custom prompts | Competitor comparison | AI traffic attribution | Content gap analysis | Crawler logs |
|---|---|---|---|---|---|---|
| Promptwatch | Yes | Yes | Yes | Yes | Yes | Yes |
| Otterly.AI | Limited | Yes | Yes | No | No | No |
| Peec AI | Limited | Yes | Yes | No | No | No |
| Profound | Yes | Yes | Yes | No | No | No |
| AthenaHQ | Yes | Yes | Yes | No | No | No |
| Semrush | Limited | No (fixed) | Yes | No | No | No |
| SE Ranking | Limited | Yes | Yes | No | No | No |
The pattern is clear: most tools handle monitoring reasonably well. The gap is in what happens after you see the data.
Building the dashboard layer
Your tracking tool collects the data. Your dashboard makes it usable. These are two different things, and conflating them is why most AI citation dashboards end up as screenshot graveyards.
Option 1: Use your tool's native dashboard
If you're using a platform like Promptwatch, the native dashboard already handles most of what you need -- citation trends, competitor heatmaps, page-level attribution, and prompt-level breakdowns. For most teams, this is enough.
The limitation: native dashboards are hard to combine with other data sources. If you want to show AI citation trends alongside organic traffic, paid spend, or revenue data, you'll need to export or integrate.
Option 2: Looker Studio
Looker Studio (formerly Google Data Studio) is the most common choice for building a custom reporting layer. It's free, connects to Google Search Console and Google Analytics natively, and most AI visibility tools either have a direct connector or can push data via API or CSV export.
A basic Looker Studio AI citation dashboard should have:
- A citation rate trend line, broken down by AI model
- A share of voice comparison chart (your brand vs. top 3 competitors)
- A table of your top 20 prompts ranked by citation rate
- A table of your top cited URLs with citation count and traffic attribution
- A "gap" section showing prompts where competitors are cited but you're not
Option 3: Airtable or Notion
For smaller teams or agencies that want something more flexible, Airtable works well as a citation tracking database. You can log weekly citation snapshots, tag prompts by topic or funnel stage, and build views that filter by AI model or competitor.
The downside is that Airtable requires more manual data entry unless you set up automation via Zapier or Make to pull from your tracking tool's API.
Setting up your prompt library
Your dashboard is only as good as the prompts you're tracking. Most teams start with too few prompts (just brand name queries) or too many (hundreds of generic category terms that are impossible to act on).
A practical prompt library for a mid-size brand typically has three tiers:
Brand prompts -- direct queries about your company. "What is [Brand]?", "Is [Brand] good for X?", "How does [Brand] compare to [Competitor]?". These tell you how AI models perceive and describe you.
Category prompts -- queries your target customers ask when they're in the consideration phase. "What's the best tool for X?", "How do I solve Y problem?", "What are the top options for Z?". These are where share of voice matters most.
Competitor prompts -- queries that explicitly mention competitors. "What's better than [Competitor]?", "[Competitor] alternatives". These surface gaps where you should be cited but aren't.
Start with 30-50 prompts across these three tiers. Track them weekly. Add new prompts when you publish new content or enter new topic areas.
The weekly review routine (90 minutes)
Consistency matters more than sophistication here. A 90-minute weekly review, done every week, will compound into meaningful visibility improvements over a quarter. Here's how to structure it.
Minutes 0-15: check the numbers
Pull up your dashboard and look at the week-over-week changes in:
- Overall citation rate (up or down, and by how much)
- Citation rate by AI model (which platforms improved, which dropped)
- Share of voice vs. top competitors
You're not analyzing yet -- just orienting. Flag anything that moved more than 5 percentage points in either direction.
Minutes 15-35: dig into URL-level data
This is where most teams skip, and it's the most valuable part. Look at:
- Which URLs were cited most this week
- Which URLs lost citations compared to last week
- Which competitor URLs are appearing for prompts where you should be visible
For any prompt where a competitor URL is outperforming yours, open both pages and compare them. Is their page more comprehensive? Does it directly answer the question in the prompt? Is it structured differently? You're looking for a specific, actionable reason -- not a vague observation.
Minutes 35-55: check crawler logs and indexing
If your tool provides AI crawler logs, spend 10-15 minutes here. Look at:
- Which AI crawlers visited your site this week (Googlebot-Extended, GPTBot, ClaudeBot, PerplexityBot, etc.)
- Which pages they crawled and how often
- Any crawl errors or pages that were blocked
A page that AI crawlers aren't visiting won't get cited, regardless of how good the content is. Crawler log data turns invisible indexing problems into fixable technical issues.
Minutes 55-75: identify one content action
Based on what you found in the URL-level review, identify one specific content action for the week. This could be:
- A new article targeting a prompt where you have zero visibility
- An update to an existing page that's being cited but losing to a competitor
- A new FAQ section on a product page to directly answer a high-volume prompt
One action per week is realistic. Two is ambitious. Ten is a fantasy that leads to nothing getting done.
Minutes 75-90: update your tracking log
Log the week's key numbers in your tracking spreadsheet or Airtable base. Note the content action you've committed to. In four weeks, you'll be able to see whether that action moved the needle.
Connecting citations to revenue
This is the part most guides skip because it's genuinely hard. But if you can't connect AI citations to business outcomes, the whole program is at risk when budgets get tight.
There are three approaches, roughly in order of accuracy:
Google Search Console integration -- GSC now shows some AI-referred traffic in the "Search type" breakdown. It's incomplete, but it's a start. Connect your AI visibility tool to GSC data and look for correlation between citation rate increases and traffic changes for specific pages.
UTM tracking on cited pages -- Some AI platforms (particularly Perplexity) pass referral data that you can capture with standard UTM parameters. Set up a filter in GA4 for referral / perplexity.ai and similar sources to isolate AI-referred sessions.
Server log analysis -- The most complete picture. Your server logs show every request, including the user agent. You can identify sessions that came from AI-adjacent referrers and cross-reference with citation data. It's more technical to set up, but it closes the attribution loop properly.
Common mistakes to avoid
Tracking too many prompts without prioritizing them. Volume estimates matter. A prompt that gets asked 50,000 times a month is worth 10x more attention than one asked 500 times. Use prompt volume data (available in tools like Promptwatch) to weight your effort.
Treating all AI models the same. ChatGPT and Perplexity have very different citation behaviors. Perplexity cites sources aggressively and frequently. ChatGPT's citation behavior varies by query type. Google AI Overviews have their own logic tied to traditional search signals. A strategy that works for one won't automatically work for another.
Ignoring Reddit and YouTube. AI models frequently cite Reddit threads and YouTube videos as sources, especially for product comparisons and "best of" queries. If your brand isn't present in those channels, you're missing a significant citation vector. Some tools (Promptwatch among them) surface Reddit and YouTube citations specifically so you can see where these third-party sources are influencing AI responses about your category.
Measuring citations without measuring traffic. Citation rate is a leading indicator. Traffic and revenue are the lagging indicators that actually matter. Build the attribution layer early, even if it's imperfect, so you have something to point to when the program needs to justify its budget.
Putting it together
Building an AI citation tracking dashboard doesn't require a massive tech stack. At minimum, you need a tracking tool that gives you URL-level citation data, a reporting layer you'll actually look at, and a weekly routine that turns data into decisions.
The tools exist. The metrics are well-defined. The routine is straightforward. What separates teams that improve their AI visibility from teams that just monitor it is whether they close the loop -- find the gap, create the content, track the result, repeat.
That loop is the whole game in 2026.