An AI visibility audit supplies a structured strategy for search and model groups to evaluate how their model reveals up on platforms like Google’s AI Overviews, ChatGPT, and Perplexity.
It measures the place the model is talked about in AI search platforms, how usually, how precisely, and primarily based on which sources.
Right here’s a step-by-step course of for a way you audit your model’s AI visibility in the present day.
Wish to flip your AI visibility audit right into a report for stakeholders?
Obtain the AI Visibility Audit Report Template, a free useful resource to share visibility gaps, AI mentions, and alternatives for progress with stakeholders.
It consists of:
- An govt abstract format for management reporting
- Scoring tables for AI + search visibility
- Competitor insights stakeholders could also be serious about
- An motion plan and notes part to doc subsequent steps
Comply with the method under that will help you fill out every part with useful insights.
Earlier than benchmarking your AI visibility, begin by defining precisely what you’re auditing and the place.
This scope determines how constant and comparable your information will probably be throughout platforms and time durations.
Right here’s what to make clear:
- AI platforms: Which platforms will you embrace within the audit? (e.g., Google AI Overviews, Gemini, ChatGPT, Perplexity, Claude, Copilot)
- Entities: Record all of the model entities and sub-entities to research, corresponding to your organization, merchandise, authors, thought leaders, or executives.
- Areas and languages: Outline geographic or linguistic scopes, particularly in case your model operates internationally.
Itemizing this info upfront avoids inconsistent baselines when evaluating visibility throughout completely different AI fashions or reporting cycles.
For instance, there’s a distinction between solely monitoring mentions of your primary model and monitoring all linked sub-brands or entities as effectively. For Ahrefs, it might embrace:
| Entity Sort | Instance |
|---|---|
| Major model | Ahrefs |
| Product manufacturers | Website Explorer, Model Radar, AI Content material Helper |
| Proprietary options | Patches, Key phrase Intents, AI Key phrase Translator |
| Proprietary metrics | Area Ranking, Site visitors Potential |
| Private manufacturers | Tim Soulo, Patrick Stox, Ryan Regulation, Glen Allsopp |
| Sub-brands | Yep, Detailed web optimization |
As soon as your scope is outlined, you’ll have a transparent basis for benchmarking and a repeatable construction you should utilize for future audits.
The following step is to evaluate your present model visibility in search.
You’ll be able to manually test your model’s search visibility without spending a dime by:
- Asking comparable questions as your prospects on ChatGPT, Perplexity, and Google.
- Evaluating whether or not your model seems within the AI responses, what’s cited, and the way outcomes differ.
- Checking month-to-month or quarterly to trace visibility modifications.
Nonetheless, guide checks can’t scale and are simply skewed by mannequin updates or personalised responses.
For constant, comparable information, you should utilize a device like Ahrefs’ Model Radar to trace visibility throughout a number of AI platforms in a single go. Enter your model and choose the area you’d prefer to obtain information for:
Pay attention to how metrics like your model’s mentions, citations, impressions, and Share of Voice change throughout numerous AI search platforms:
- Mentions: The variety of instances your model is talked about in AI responses.
- Citations: The variety of instances your model’s web site is cited as a supply in AI responses.
- Impressions: An estimation of publicity primarily based on how usually responses containing your model are proven in search outcomes.
- AI Share of Voice: How usually your model is talked about in AI responses in comparison with opponents.
You can even benchmark whether or not persons are looking for your model within the Search demand tab:
This graph reveals how branded search curiosity modifications over time. An increase in search quantity means extra persons are actively searching for your model, whereas progress in search question rely signifies customers are asking a wider vary of questions on you (from brand-level queries to particular product or characteristic searches).
Take a look at these metrics periodically to see how your model’s visibility modifications over time and to report the outcomes of your optimization efforts.
When you’ve gathered your benchmarks, it’s time to go deeper. Begin by taking a look at the entire AI responses that point out your model within the platforms you care about.
Navigate to the AI responses report, choosing the AI index you care most about to start your evaluation:
The responses will appear to be this within the dashboard, with bolded mentions of your model:
Strive utilizing these filters to assist make the evaluation simpler:
| Use Case | Filter Settings |
|---|---|
| See what AI says about you for branded searches | Question incorporates {your model} |
| See how AI talks about your model normally | Response incorporates {your model} |
| Analyze when AI makes use of your content material as a supply | Quotation incorporates {your area} |
You’ll be able to combine and match filters to deepen your evaluation. Right here’s a useful information with extra filter mixtures and suggestions price making an attempt out.
As you undergo the responses, you’ll begin to discover patterns about how AI mentions your model.
For instance, you would possibly discover that your model seems in lots of comparative searches, however your content material isn’t cited. Which means that the AI platform sources details about your model from opponents or third-party, impartial websites which have in contrast you to opponents.
Different patterns price taking note of embrace:
- Accuracy: Is your model info within the response correct?
- Sentiment and framing: Is your model introduced in a optimistic, impartial, or unfavorable gentle?
- Differentiation: Are any distinctive attributes and positioning of your model current within the response, or is it generic info?
- Authority: Is your model introduced because the lead point out with probably the most authority in comparison with others?
- Name-to-action: Does the response make you the beneficial supply to information the person on the subsequent step of their journey?
When you discover incorrect, outdated, or probably dangerous details about your model, make an inventory of the queries, responses, and cited sources so you possibly can overcome misinformation and model inconsistencies.
Repeat this course of for each AI search platform you wish to analyze. It’s possible you’ll discover that completely different patterns emerge for every platform.
There are two kinds of evaluation you are able to do right here. The primary is to seek for your model and filter out queries that don’t comprise your model title. Within the filter settings, set it to “Question doesn’t comprise {your model}”:
Then, analyze the Matters report back to see what your model is usually related to. You’ll be able to see the precise responses concerning the matter that additionally point out your model:
For instance, Ahrefs is carefully related to web optimization subjects corresponding to key phrase analysis, backlinks, and web site visitors. Usually, for those who’re seen in AI responses, the core subjects you wish to be aligned with will probably be mirrored right here.
However you can even comply with by way of with the subsequent step that will help you determine subjects you is perhaps lacking. As a substitute of doing a model search, this time, solely embrace the principle subjects you wish to be discovered for:
It may assist to research one matter at a time if there are a couple of you wish to take a look at. Then use the filters to take away any outcomes that point out your model within the question, response, and citations.
Take a look at the Matters report once more to seek out sub-themes you might wish to connect with your model. Search for gaps you’d like to shut, like the place your model is:
- Not talked about, however a competitor is
- Not strongly related to vital sub-topics in AI responses
- Lacking content material codecs AI ceaselessly cites (like guides, movies, opinions)
Then, you possibly can shut the gaps in your web site by both publishing new pages or enhancing the subject protection of current ones.
You should utilize AI Content material Helper for this. It analyzes your content material and helps you determine important subjects to cowl on every web page.
Analyzing your model’s top-cited pages reveals which content material AI programs contemplate most authoritative, serving to you perceive, replicate, and develop the alerts driving your visibility in AI search.
There are two methods to seek out your hottest pages which can be cited in AI responses.
The primary is to make use of a device like Model Radar, which reveals you the precise queries and responses you’re cited for. The second is to research your page-level visitors from AI through the use of a device like Ahrefs’ Internet Analytics or GA4.
Examine your top-cited pages in Ahrefs’ Model Radar
Begin typing your model in Model Radar and choose the choice so as to add variations:
This can permit you to add your web site as a part of your model entity:
You’ll be able to solely see information about citations for those who add a website to trace, so all the time double-check this setting for those who can’t see any quotation information within the dashboard.
Then, navigate to the Citations tab within the dashboard and take a look at what number of citations your model has earned throughout completely different platforms:
Clicking these values will take you to the responses that embrace citations. Nonetheless, you can even take a look at the Cited Pages report for those who simply wish to see the record of top-cited pages.
In some circumstances, you’ll see an inventory of pages out of your web site and others which can be cited in the identical responses the place your model is talked about. Use the filter to indicate solely your area:
You’ll find yourself with an inventory of top-cited pages in descending order primarily based on what number of responses used the web page as a supply: 
Cross-check the highest pages incomes AI visitors in your analytics information
The second methodology you possibly can strive is to take a look at your net analytics platform and analyze which pages are getting probably the most visitors from AI.
Though you received’t have the ability to see the precise queries and responses driving these clicks to your web site, you possibly can nonetheless get an concept of which pages really earn clicks. Simply because a web page is well-cited amongst AI responses doesn’t imply folks will click on on it.
You’ll have to have AI monitoring in your analytics device already set as much as get this information. When you use Ahrefs’ Internet Analytics, that is arrange for you robotically:
Search for an inventory of the highest pages visited most from AI platforms:
Cross-check these pages towards these in Model Radar. For every one, test the responses the web page is cited in:
This helps you see which queries and responses drive probably the most clicks from AI search outcomes to your web site.
Repeat the method throughout completely different AI platforms (e.g., AI Overviews, ChatGPT, Perplexity) to match click-through patterns.
A significant side of incomes model visibility in AI is how others discuss your model on-line. You’ll be able to run a full audit to research your model mentions all over the place folks discuss you.
Doing this will provide you with a holistic concept of the language patterns and information about your model that AI is prone to embrace in responses about you.
When you’re working agency-side or simply don’t have time to do the complete course of, check out this shortcut that can provide help to prioritize which on-line mentions to research first.
Run an empty search in Model Radar and use the filters to solely embrace the AI responses that point out your model:
Then, take a look at the Cited pages report. You’ll get an inventory of all pages that AI cites for responses that point out your model. As an illustration, Zapier and Reddit are ceaselessly cited in responses that point out Ahrefs:
You’ll be able to take a look at the responses themselves for an concept of how your model was talked about.
Then take a look at the content material cited on these web sites to see precisely if (and the way) they point out your model.
This evaluation is helpful for:
- Recognizing your greatest net mentions and the way they affect the best way AI additionally talks about you
- Figuring out misinformation in AI responses and monitoring down the precise supply
- Seeing what content material codecs are ceaselessly cited in your trade (and publishing comparable posts)
- Partnering with web sites that ceaselessly get cited and infrequently point out you favorably
- Studying how folks (and AI) place you towards your opponents
This supplies a prioritized record of net mentions that considerably impression how AI describes your model. Replicate one of the best ones for quick-win visibility positive aspects whereas additionally reaching out to those who might have updating.
When you’ve recognized these influential mentions, the subsequent step is to seek out gaps you possibly can shut or alternatives your opponents have but to faucet into.
To benchmark your model’s AI visibility towards opponents, search your model after which add a number of opponents like so:
You can even check out the AI options to auto-populate your prime opponents in AI search.
Within the Overview tab, you’ll have the ability to examine your AI visibility instantly to every competitor you entered. You’ll be able to see every model’s visibility metrics like mentions, impressions, citations, and AI Share of Voice for every AI search platform.
Hover over the platforms you care most about till you see this window pop up:
Every choice will take you to a pre-filtered view of the AI responses report.
- Solely model: This report will present you queries and AI solutions the place the one model talked about is the one you’ve chosen. It’s nice for assessing brand-specific queries and responses.
- With others: This report reveals queries and/or responses that comprise all of the manufacturers you’re evaluating. It’s nice for assessing subjects which can be prone to point out you alongside opponents.
- Others solely: This report reveals queries that solely point out the opposite manufacturers you’re evaluating towards. It’s nice for locating subjects that point out your opponents however not you.
Take a look at these studies periodically for insights into which opponents is perhaps rising their AI visibility shortly.
When you add a subject as a part of the principle search, you can even hone in on any particular subjects the place opponents are talked about however your model isn’t.
Take note of the AI Share of Voice metrics to see who has the best visibility share in responses about that matter.
It’s additionally price testing the Cited domains and Cited pages studies to determine the highest sources mentioning every model.
Search for web sites that point out opponents greater than you and see how one can enhance your mentions inside them, too.
There are quite a few prospects right here, relying on which aggressive insights matter most to your stakeholders.
By this level, you’ll have a transparent image of how your model seems in AI search: what’s working, the place you’re absent, and the way you examine to opponents. The ultimate step is popping these insights into motion.
Use your findings to prioritize subsequent steps primarily based on impression and energy:
- Repair: Appropriate misinformation or outdated model descriptions
- Construct: Increase content material protection for lacking subjects and codecs that AI cites most
- Affect: Strengthen relationships with authoritative websites that AI depends on for citations
Relying on the duty, you might have to collaborate with different groups that will help you implement these actions. For instance, in case you have a PR staff, they are able to attain out to publications and proper misinformation concerning the model in different folks’s content material.
If you must search approval earlier than you can begin implementing these actions, take the time to create a report to your stakeholders.
Obtain our free AI Visibility Audit Report Template for those who don’t have already got one. Substitute all of the pink sections and summarize your key findings in scorecards and fast snapshots of key factors.
Lastly, flip your audit into an ongoing cycle, ideally month-to-month in case you have the bandwidth. In any other case, quarterly will do.
With AI search platforms altering as shortly as they’re, and opponents scrambling to win the most important share of voice, it’s price checking your visibility ceaselessly.
Document the identical metrics as you probably did in step two so you possibly can see how your visibility shifts over time.
Closing ideas
An AI visibility audit tracks how your model is introduced in AI search over time. It measures the quantity of responses that point out your model title or cite your web site as a supply. Nonetheless, it additionally assesses qualitative alerts like how your model is represented.
To enhance your visibility in AI search, take a look at the next guides:
