Triangulating your brand’s impact in the “New Normal” of AI

The onset of text-based generative AI is repeatedly raising a new question: how can you measure AI’s impact on marketing? With new tools launching left and right, plus an AI ecosystem that can literally evolve overnight, attempting to answer that question is becoming more complex, not less.

Here’s the reality: there’s no single metric that will unlock meaningful answers about your brand’s AI presence — but that doesn’t mean it can’t be analyzed. Instead, measuring AI effectiveness requires triangulation. You’ll need perspectives from multiple angles before meaningful insights emerge. 

In the midst of this chaos, let’s take a step back and consider what we truly care about when it comes to tracking how our brands are showing up in large language models (LLMs). We’ll explore three complimentary approaches that, can inform real, actionable insights when viewed holistically: visibility, site referrals, and sentiment.

Visibility — mention equity and competitor impact

The question you’re trying to answer here is straightforward: “How likely is my brand to show up in AI results compared to my competitors’? Just as search engine optimization (SEO) aims to place your brand at the top of search results, the goal of answer engine optimization (AEO) or generative engine optimization (GEO) is to “rank” highly within AI responses. 

An easy place to start is by counting the number of times your brand is mentioned and comparing that to the number of times your competitors show up for a given set of prompts.

Let’s say you have 10 related prompts you want to rank highly in. After running them in an LLM, you find:

• Your brand was mentioned 10 times

• Competitor A was mentioned 25 times

• Competitor C was mentioned 5 times

With this data, you can infer that your brand owns about 25% of the mention equity — the percentage of references to your brand versus your competitors — within the context of your chosen prompts for that LLM. You can think of it as a measure of how much “space” your brand owns. The more data you’re able to collect, the more accurate this will be. 

This simplified exercise illustrates the way many AEO/GEO reporting tools work; they just do it at a much larger scale and with far more automation.

Choosing the right prompts is a critical step. Unlike SEO, where you can peek at Google Search Console or search trends to get near real-time query data, there’s currently no way of knowing the actual prompts users are running in LLMs out in the real world. Instead, we make educated guesses based on high-intent, high-value questions that closely resemble real user behavior. (Pro tip: using AI itself can be helpful for brainstorming new prompt ideas if you’re stuck!)

Site Referrals — AI as a traffic driver

When a user clicks a link in an LLM response, the AI platform often includes an urchin tracking module (UTM) in the link — a bit of extra text at the end that tells your web analytics software where the user came from. 

This makes it surprisingly trivial to track that user’s activity if they hit your website. In fact, some web analytics tools, such as Google Analytics 4, are coming pre-packaged with an AI search channel bucket. Others may require building out a custom channel grouping.

If you examine the source field in your traffic reports, you’ll notice AI referrals identified by domains like chatgpt.com, claude.ai, and others. By grouping these sources into a single AI channel, you can quickly assess:

• How much traffic is driven by AI

• What meaningful actions users perform when they hit your website

• What landing pages they’re being directed to most often

In web analytics terms, a “landing page” is the first page reached when a user starts a session. It can be virtually any page on a website. Landing pages matter in the context of AI search reporting because they help clarify exactly which URLs on your site are cited by LLMs. This provides insights into what content is performing well and uncovers where your site may have topical gaps.

Sentiment — how AI influences the narrative

Take the cliché “all press is good press”, bury it in a deep hole, and leave it there forever. When customers talk about your brand, you want them to be positive rather than neutral, or worse, making negative comments that steer others away. 

The same is true with AI. People routinely go to AI with questions like “Which of these products should I buy?”, “What are the pros and cons of ABC product?”, and “Is XYZ company a good place to work?” Being in AI’s good graces is just as important as being cited, if not more so. 

Tools such as Meltwater’s GenAI Lens can help you automate sentiment tracking. When negative responses crop up, take a close look at what’s being said about your brand. Is the information accurate? Can you find the same sentiment being expressed elsewhere online? Is the LLM picking up on an angry comment, an outdated news piece, or hallucinating the whole thing? Regardless of the source, sentiment analysis can highlight areas where PR, content, or reputation management efforts may need reinforcement.

Building a more complete picture

Visibility, site referrals, and sentiment are just three of the many angles you can look to when evaluating your marketing impacts on LLMs. Alone, each should be considered as part of a larger puzzle. The more pieces you can put together, the clearer the picture will be. Here’s a few more tips to get the most from these reporting exercises:

• Monitor multiple AI platforms. Each one works in its own unique way and can yield different — often surprising — results. Currently, the most relevant platforms include ChatGPT, Claude, Gemini, Copilot, Deepseek, and Llama.

• Remember that LLM responses are inherently unpredictable and will vary from user to user. Variables like chat history, precise wording of natural-language queries, and the baked-in randomness of LLMs make it impossible to predict with 100% certainty what responses will look like for your entire audience.

• Understand response types. Some AI answers are pre-trained, while others rely on live retrieval via search engines. Optimization strategies can differ meaningfully between the two.

• Prioritize qualitative insights over quantitative metrics. Metrics matter, but the end goal of tracking should be to reconstruct what’s happening in the real world. Make sure any data you report is supported with analysis and an interpretative lens.

The AI search landscape is evolving rapidly. Tools will improve, methodologies will mature, and best practices will shift. Flexibility is essential. But regardless of where this road takes us, the evergreen truth is that the more angles we examine, the wider our perspective will become and the deeper our insights will be. 

So, the next time you look for answers to AI’s impact on marketing, consider a triangulation approach. You may find your reporting becomes far more meaningful and actionable than you ever expected.