The 5-Step AI Visibility Audit You Can Do Right Now

People’s search habits are changing. AI engines like ChatGPT and Gemini are now part of the consumer discovery process.

For PR teams well-versed in capturing conservations through media monitoring and social listening, these AI tools present a new challenge: understanding how their brands appear across LLM (large language model) responses.

Navigating ChatGPT prompts and answers may seem like a daunting task, so that’s why we’ve put together a simple five-step AI visibility audit to find out where your brand stands. This process will help give you insight into how and where your brand is mentioned (or not) and the tone of the responses.

Step 1: Identify questions people are asking about your brand

Start by thinking like a consumer. What questions are they asking that could lead them to your brand? Also, what questions could they be asking directly about your brand?

Think in both broad and specific terms. For example, if people are interested in purchasing an EV, there are many things that feed into their buying decision. Research across Google’s “People Also Ask,” Reddit communities, and Quora to surface real questions being asked around the topic. You should also plug in searches to your media monitoring and social listening tools to uncover more discussion points.

According to Matthieu Danielou, co-founder and CEO of Search Intelligence platform Trajaan, having a “master prompt” with added prompts for context gives a more complete picture of a topic on an LLM.

In the example of electric vehicles, he suggests four questions to surface a range of responses:

• What are the best EVs?
• What is the best EV for a family?
• What is the best EV for a single person?
• What is the best EV for a single person living in London?

Danielou adds: “[With LLMs] we simulate different personas to get a sense of all the answers that they would get and to find some discrepancies, some differences between them.”

These prompts cover awareness, direct brand comparisons, and recommendations – all themes that are likely to produce brand mentions on AI platforms.

Remember that recent research has shown that AI chatbots are in the habit of people-pleasing, so try and keep your questions direct and as neutral as possible to avoid confirmation bias.

Step 2: Run your prompts and monitor responses

With your questions finalized, pick one AI search engine to start your audit. ChatGPT, Claude, Gemini, or Perplexity are good places to begin. Run each of your questions manually and monitor what comes back.

Repeat these identical prompts across more AI engines. You'll likely find meaningful differences in how each one frames your brand, which competitors it surfaces, and the sources it cites.

Remember that AI responses aren't static. The same audit conducted on different days can return different results. This variability makes a manual audit over a longer timeframe labor-intensive. That’s why it’s worth starting small for an initial visibility audit. If you’re keen to scale up, platforms like Trajaan can monitor GenAI tools with hundreds of prompts simultaneously.

Step 3: Build a scorecard

Document your results in a scorecard that makes patterns easy to identify.

Ask yourself questions like: Does your brand appear? Is it in a positive, neutral, or negative tone? Who else is being mentioned alongside you? Is the information about your brand accurate?

Here’s an example scorecard of AI brand monitoring, using our previous EV questions:

Once you’ve applied a scorecard to your own prompts, you'll begin to see where you have a consistent presence, where gaps exist, and which competitors are winning share of voice in AI results.

Step 4: Identify your wins and weaknesses

Look at your findings more closely to find your big wins. Which LLMs discuss your brand favorably? Do they reflect recent brand positioning or campaign work? These responses will signal that your existing messaging – through earned media, press releases, or owned content – is working and being picked up by AI.

Next, look at the gaps. Are there prompts where your competitors are named but you're not? If you need to be in that space, view the absence as a content opportunity. If your brand isn't part of the conversation around a particular topic, that's a signal to build more credible content and messaging around it.

Lastly, check for inaccuracies. If any responses are describing your brand incorrectly, or drawing on outdated information, that’s a reputational risk worth addressing.

Step 5: Bring your findings into your PR plan

The results of your audit should be used to help shape your ongoing PR plan. LLMs draw from a wide variety of sources when generating responses, many of which PR teams either control or influence. That means your press releases, earned media, and social media activity all play a role in how AI views your brand.

With that in mind, it’s worth considering which PR levers you can pull to address what the audit has surfaced. You might consider:

• An earned media push: If competitors are dominating category conversations, your brand's key differentiators can help shift the balance over time.
• Media outreach and refreshing owned content: If there are factual inaccuracies in AI responses, this combination can help introduce corrected information into the sources LLMs learn from.
• Press releases and social content: If your brand is barely visible at all, a consistent drumbeat of content around specific products or key topics can start to build presence.

None of these are instant fixes, and some AI platforms handle content differently. For example, AI Overviews on Google use live, real-time data pulled at the time of search (powered by the Gemini model).

ChatGPT, however, can pull live data but doesn’t rely on it solely for every answer. The response depends on the question asked. (You can often see when it’s using live data by the “searching the web” message in the chat.)

All this means that your PR efforts may take some time to appear across LLMs. A press release you published last week may not shape LLM responses for weeks or months. This is where it pays to keep revisiting your brand’s AI visibility and making search intelligence part of your wider PR strategy.

The bottom line

This five-step audit is a practical starting point for any PR team that wants to understand how their brand is showing up in the AI-driven discovery process. It's manual and has limitations, but it gives you a baseline to work from.

The challenge is doing this consistently and at scale. Running four prompts across three platforms once a quarter is manageable. Monitoring dozens of questions across a fast-moving competitive ecosystem is not. 

That's where a dedicated tool makes the difference. Cision's Search Intelligence platform, Trajaan, is built to do exactly that, allowing you to monitor how your brand, your competitors, and your industry are represented across AI search engines.

Alongside the tools and tactics, there's also an important mindset shift at play. AI has now joined journalists and consumers as an audience member at the heart of public relations. With LLMs increasingly the first place people turn to when researching a brand, PR teams need to start treating AI search as a key communications channel in its own right.

Danielou notes that brand teams need to be able to work “in a manner that will influence the answers of these LLMs.”

“What really matters is to make sure that you push the right messages to this new audience, to this new LLM persona,” he says. “You need to do it by leveraging all the right channels, your websites, your media partners, your social channels – so that at the end of the day, you own the moment, you own the answers.”

Find out what GenAI search engines are saying about your brand and industry. Explore Cision’s Search Intelligence solution.