How businesses can
monitor AI

When Google became the dominant search engine, the first thing smart businesses did was learn to measure their rankings. Google Search Console, Analytics, rank trackers. An entire industry emerged around measuring search performance. AI search is at that same inflection point right now. 51% of UK adults already use AI search tools. 30% of UK consumers research products through AI chatbots. But almost no businesses are monitoring what AI says about them.

The problem with not monitoring is straightforward: you cannot improve what you do not measure. Your competitors might already be getting recommended by ChatGPT while you are invisible. A potential customer might ask Gemini "Who is a good accountant in Birmingham?" and get three names back. None of them yours. Without monitoring, you would never know this is happening. You would just notice that leads are declining and have no idea why.

AI monitoring is also fundamentally different from Google monitoring. With Google, your rankings are relatively stable. If you rank number 3 for a keyword today, you will likely rank number 3 tomorrow. With AI, responses vary between sessions. Ask ChatGPT the same question five times and you might get five different answers. The businesses mentioned change. The order changes. The sentiment changes. This variability makes single manual checks unreliable. You need systematic, repeated monitoring to get an accurate picture of your AI visibility. Learn more about why responses differ in our article about why AI results differ between platforms.

Another reason monitoring matters: AI platforms pull from different sources. ChatGPT uses training data plus live web search. Gemini draws from Google's index and Google Business Profile data. Perplexity relies heavily on recent web content. Claude emphasises authoritative sources. Your business might be visible on Gemini because of your strong Google Business Profile but completely invisible on ChatGPT because your website lacks structured data. Without cross-platform monitoring, you cannot identify these gaps.

The competitive intelligence dimension is equally important. AI monitoring does not just show you whether you are mentioned. It shows you who is mentioned instead. If a customer asks "Who is the best estate agent in Manchester?" and ChatGPT recommends three of your competitors but not you, that is actionable intelligence. You can analyse what those competitors are doing differently. Are they on more review platforms? Do they have better structured data? Do they publish more FAQ content? The answers point you toward specific improvements you can make.

Before investing in tools, you can get a baseline understanding of your AI visibility through manual testing. This is not a replacement for systematic monitoring, but it gives you a starting point and helps you understand the landscape. Here is how to run a basic manual audit.

Step 1: Define your test queries

Write down 10 to 15 questions that a potential customer would ask. Mix generic queries ("best solicitor in Manchester") with specific ones ("how much does conveyancing cost in Manchester") and comparison queries ("solicitors vs conveyancers for house purchase"). Include your location and service variations.

Step 2: Test across platforms

Ask each query on ChatGPT, Gemini and Perplexity. Use new chat sessions for each query to avoid context contamination. Note down whether your business is mentioned, in what position, and what AI says about you. Also note which competitors are mentioned.

Step 3: Repeat after two weeks

AI responses change over time. Run the same test again after two weeks to see if your results are consistent or fluctuating. If they fluctuate significantly, single checks are unreliable and you need systematic monitoring. Keep a simple spreadsheet with date, platform, query, mentioned (yes/no), position and sentiment.

Manual monitoring has clear limitations. It is time-consuming. You can only test a small number of queries. AI responses vary between sessions, so your results might not represent what customers actually see. You cannot track trends over time with any statistical confidence. And you cannot monitor continuously without making it a full-time job.

For most businesses, manual monitoring is a useful starting exercise that reveals whether you have an AI visibility problem. Once you confirm that problem exists (and for most UK businesses, it does), you need automated monitoring to track your progress consistently. Think of it like the difference between manually checking your Google rankings once a month versus using a rank tracker that monitors daily.

Monitoring is only valuable if you act on what it reveals. Here are the most common patterns you will discover and what to do about each one.

Pattern 1: Not mentioned anywhere

If AI never mentions your business on any platform, your online presence needs fundamental strengthening. Start with the basics: complete your Google Business Profile, set up Trustpilot and LinkedIn profiles, add schema markup to your website and publish FAQ content answering your most common customer questions. These foundational steps give AI platforms the signals they need to discover and recommend your business. Read our complete guide on how to get visible in ChatGPT for a detailed action plan.

Pattern 2: Mentioned on some platforms but not others

This is the most common pattern. A business visible on Gemini (because of a strong Google Business Profile) but invisible on ChatGPT (because the website lacks structured content). Look at what the platform where you are visible is pulling from, then replicate those signals for the platforms where you are not. Gemini favours Google Business Profile data. ChatGPT favours well-structured websites and review platforms. Perplexity favours recent, citable content.

Pattern 3: Mentioned but with negative sentiment

If AI mentions your business but with qualifications or negative context, look at your review profiles. AI pulls sentiment from reviews, and a pattern of complaints about slow service, poor communication or pricing issues will show up in AI recommendations. Address the underlying issues first, then actively seek positive reviews to shift the balance. Responding professionally to negative reviews also helps, as AI picks up on businesses that engage with feedback.

Pattern 4: Mentioned but competitors are mentioned more

Analyse what your competitors are doing differently. Common advantages include: more reviews (especially recent ones), stronger platform presence (listed on more directories), better structured data, more comprehensive FAQ content and more third-party mentions (press coverage, industry articles). Pick the gap that is easiest to close first and work systematically through the others. Our article on why AI mentions your competitors gives a detailed breakdown of competitive AI advantages.

Pattern 5: Visibility fluctuates significantly

Some fluctuation is normal because AI responses vary between sessions. But if your visibility swings dramatically (mentioned 80% one week, 20% the next), it suggests your signals are borderline. AI is not confident enough to consistently recommend you. The fix is to strengthen all your signals: more reviews, more platform presence, better content, stronger structured data. The goal is to make your business an obvious recommendation that AI consistently includes.

Setting up a monitoring cadence

For most UK SMEs, a weekly monitoring review is sufficient. Check your key metrics (mention rate, position, sentiment, competitor comparison) each week and note trends. Make improvements based on the patterns you observe. Schedule a deeper monthly review where you analyse which specific changes correlated with visibility improvements. This cadence gives you enough data to identify trends without becoming overwhelmed by noise. Track your progress with a tool that gives you historical data so you can see the trajectory over months, not just snapshots.