A segment defined by demographics, interests, or query patterns used to track brand appearance in AI search responses.
Marketing professionals concerned with optimizing content for AI discoverability and tracking brand visibility.
01What it is and how it works
A Custom Audience is a set of criteria you define—such as age range, interests, location, or typical search topics—that the product uses to filter AI search results or simulate queries. The product then analyzes AI-generated responses (e.g., from ChatGPT, Bard, or Perplexity) for brand mentions within that audience’s context. For example, if you define a Custom Audience of “fitness enthusiasts aged 25–40,” the tool will run queries related to fitness and check if your brand appears in the AI’s answers. The mechanism relies on the product’s ability to either (a) filter existing AI search data by audience attributes or (b) generate synthetic queries that match the audience’s likely interests. This allows you to see brand visibility not just overall, but for the segments that matter most to your marketing strategy.
You pick a group of people (like “tech enthusiasts” or “parents of teens”) and the tool measures how often your brand shows up when AI answers questions from that group.
02What to do about it
Start by mapping your buyer personas to Custom Audiences. For each persona, list the topics they ask AI about and the language they use. Then create an audience in the product using those criteria. Set up alerts for when brand mention rates drop below a threshold for a specific audience. Review audience performance weekly: compare share of voice across audiences to identify which segments have the weakest brand presence. Adjust your content strategy to target those gaps—for instance, if your brand is missing in responses about “budget travel” for the “young families” audience, create content that answers those queries and optimize it for AI discoverability.
03How it is measured or noticed
The key metric is brand mention rate per Custom Audience: the percentage of AI responses that include your brand when queries match that audience’s criteria. You can also track share of voice (your brand’s mentions vs. competitors’ within the same audience), sentiment of mentions (positive, neutral, negative), and the topics that trigger mentions. The product’s dashboard shows these metrics over time, broken down by audience. A sudden drop in mention rate for a previously strong audience signals a change in AI behavior or a competitor gaining ground.
04Common mistakes
- Defining audiences too broadly (e.g., “all adults”)—this dilutes insights and misses segment-specific trends.
- Using static criteria that never update—audience interests shift, and stale definitions produce irrelevant data.
- Ignoring the AI’s context window: a Custom Audience based on demographics alone may not reflect the actual queries users ask.
- Overlapping audiences that double-count the same mentions, leading to inflated metrics.
- Assuming a Custom Audience matches your real-world customers—validate with survey data or CRM segments.
05Limits
Custom Audiences work best when the audience is large enough to generate statistically significant data. For very niche segments (e.g., “left-handed violinists in Montana”), the product may have too few queries to produce reliable metrics. Also, Custom Audiences are not the same as lookalike audiences—they are based on explicit criteria, not algorithmic similarity. They do not measure actual user behavior, only simulated or filtered AI responses. Finally, the product cannot guarantee that the AI’s answers reflect the real-world opinions of that audience; it only shows what the AI generates when prompted with audience-relevant queries.
06A worked example
I defined a Custom Audience called “tech early adopters” with criteria: age 22–35, interests in AI, gadgets, and startups, location urban. The product ran 200 queries about “best smart home devices” and “AI productivity tools.” My brand appeared in 12% of responses for this audience, compared to 8% for the general audience. I saw that competitors were mentioned more often in queries about “affordable AI tools,” so I created content targeting that topic and improved my mention rate to 18% within two weeks.
Frequently asked questions
How is a Custom Audience different from a regular audience segment in analytics?
A Custom Audience in this product is specifically designed to track brand mentions in AI search responses, not general user behavior. It uses AI query simulation or filtering based on defined criteria, whereas standard segments rely on historical data from clicks or page views.
Should I create Custom Audiences for every buyer persona?
It depends on whether each persona represents a large enough group to yield statistically significant brand mention rates. Start with your most important personas and expand only when you have sufficient query volume for reliable data.
How do I actually set up a Custom Audience in the product?
You define criteria such as age range, location, interests, or typical search topics. The product then either filters existing AI search results to match that profile or simulates queries from that audience's perspective to isolate brand mentions.
Can a Custom Audience give reliable data if my brand is niche?
It works best when the audience is large enough to generate statistically significant data. For very niche audiences, you may need to broaden criteria or combine segments to avoid inconclusive results.
What happens if I define a Custom Audience too narrowly?
You risk getting insufficient data to measure brand mention rate reliably, leading to inconclusive or misleading insights. You might miss broader trends and waste effort on a segment that doesn't produce actionable results.
How long does it take to see results after defining a Custom Audience?
Results appear as soon as the product runs queries or filters data for that audience. However, meaningful trends may require a few days of data collection depending on query volume and the specificity of your criteria.
Asked out loud
spoken, not typedThe same term in the words somebody uses speaking to an assistant rather than typing into a box — written from the situation, which is why each one carries the situation it came from.
Yes, you can create a Custom Audience with age range and interest in internships. The product will simulate queries from that perspective and show the brand mention rate for that segment.
Yes, you can set up a Custom Audience with location and business-related interests. It will start tracking immediately, and you can review the results later when you're free.
It's possible. Custom Audiences need a large enough base for statistically significant data. If your audience is too niche, you might get sparse results. Try broadening the criteria to include related interests.