- +44(0)28 9099 8390
Measuring the impact of being recommended by LLMs like ChatGPT, Gemini, and Claude is possible when approached correctly. This comprehensive AEO measurement guide explores the limitations of traditional marketing measurement and outlines what marketing teams can do to effectively measure the impact of AI tools on their business.
Published: 8 July 2026
Author: Adam Grant
Traditional search was simplistic to measure. You either drive traffic via organic search (SEO), or you pay for a Google Ad. You can then project the number of leads based on search traffic, CTR, and conversion rate. Now that buyers shift to using AI search, traditional measurement no longer effectively captures impact.
| AI Search | Traditional Search |
|---|---|
|
Limited Prompt Consolidation
Unlimited variance in prompts. Buyers phrase the same question in thousands of different ways, shaped by the conversation they’ve had with the LLM up to that point.
|
Clear Search Volumes
Keywords can be researched and ranked against known, measurable demand.
|
|
Multiple AI Tools
Different AI tools that cite different sources resulting in variance in answers.
|
One Search Engine (Google)
Optimising for a single platform covers the majority of the market.
|
|
Prompt Variance
Two prompts run right after each other may give a different answer, as model outputs vary by design.
|
Consistent Rankings
Google rankings have little to no variance on a day-to-day basis. The business ranks first from one search to the next.
|
|
Very Limited Attribution
AI tools rarely pass referral data, so visits influenced by an AI recommendation often arrive as direct or organic traffic.
|
Clean Attribution
Click-throughs, sessions, and conversions can be tracked end-to-end through standard analytics.
|
|
Additional Value
AI tools don’t just help with discovery. They actively evaluate options and recommend a solution, resulting in high-quality leads.
|
Discovery Only
Traditional search helped you get discovered. Your website and sales team would need to convert this traffic.
|
The B2B buyer journey is long and complex, particularly in the technology sector, where buying committees typically span IT, management, and finance. These buyers use AI tools to discover, research, and compare potential solutions. The challenge is that buyers have multiple touchpoints with marketing, and trying to reduce that journey into one touchpoint will inaccurately represent the impact of different marketing tactics.
If your business is recommended inside an AI tool and the buyer then searches for you on Google and lands on your site that way, the session is attributed to organic search, not the AI tool that surfaced you in the first place.
If you need new software, you typically start with a problem. Your finance team is drowning in manual invoice processing, and you need to automate it. You then go back and forth with an LLM before asking for a recommendation for AP automation software. That prompt and query fan-out didn’t search a clean search term like in Google. It took the context from the conversation and created query fan-outs to answer the user’s question as best it could. So how can you expect an AI visibility tool like Profound to capture these prompts? These tools will track the prompts you provide, but what about the variance of these prompts? It’s like tracking 100 long-tail keywords without keyword volume, when there are 10,000 other ways someone may be searching for a solution.
Buyers are using different LLMs to research and discover potential providers, which makes tracking across multiple platforms essential. ChatGPT currently leads with around 68% usage share, down from 87% in previous years as smaller models rise for different use cases. Gemini has grown to roughly 18%, with DeepSeek, Grok, Claude, Perplexity, and Copilot remaining in single digits but each growing steadily. Without tracking across these tools, businesses miss large segments of the market and end up with unreliable data.
Focus your tracking on bottom-funnel commercial prompts. These act as a proxy for the wider cluster of prompts a buyer might use, where the variations are too numerous to track individually. In the MSP space, that means prompts like “best MSP for law firms”, “top MSPs for mid-market companies”, or “best MSP for cloud migration”.
Useful structures to test against your own market include “what are the best [solution type] providers for [your market]?”, “which companies specialise in [your core service] for [your buyer type]?”, and direct comparisons like “compare [your business] with [competitor]”. For each prompt, record which AI tools include you, where you appear relative to competitors, and what the answer says about you.
A single AI response is not a reliable signal. Run the same prompt twice and the recommended providers can shift, so a once-a-day check tells you very little about your actual visibility.
The fix is to sample at scale. Run the same prompt ten times in a single session and track how often your brand appears, alongside which competitors show up. Repeated weekly, this gives a far more accurate read on whether visibility is genuinely moving.
The volume and quality of citations earned in AI answers is one of the clearest signs an AEO programme is gaining ground. Citations move before commercial impact does, which makes them a strong leading indicator.
Track citations by LLM as well as in aggregate. The split tells you where your programme is performing best and which sources each LLM is leaning on, both of which feed directly back into content and outreach priorities.
The authority signals that feed those citations are largely external, so it’s worth tracking the inputs alongside the outputs. None of this requires complex tooling. A simple tracking document updated monthly produces a clear picture of whether the authority footprint is growing in the right direction.
Buyers are not loyal to one tool. ChatGPT still holds the largest share, but a meaningful portion of research is happening on Gemini, Claude, Perplexity, Copilot, and Grok, so measuring ChatGPT alone misses close to half of buyer activity.
We recommend tracking across ChatGPT, Gemini, Claude, Google AI Mode, Google AI Overviews, Perplexity, and Copilot. As of July 2026, overall usage is as follows: ChatGPT holds 52.1%, Claude 21.5%, Gemini 13.3%, Copilot 8.4%, Perplexity 3.4%, and Grok 0.6%. Perplexity and Gemini both punch above their weight in the B2B tech sector, as we have seen just as many leads from these tools as ChatGPT.
Three sources need to be read together. LLM referral traffic is the most direct signal: ChatGPT, Perplexity, and others pass referral data, and while volumes are still modest, a rising trend line is one of the cleanest indicators that AEO work is reaching buyers.
Direct traffic is where most AI-influenced visits actually land. A buyer reads an AI answer, sees your name, and types your domain in later. None of that shows up as a referral, but a sustained lift in direct traffic that tracks alongside rising visibility is a strong corroborating signal.
Organic traffic matters too, as the line between SEO and AEO is blurring. Google AI Overviews and AI Mode now sit on top of organic results, so movement here reflects both traditional search performance and AI-driven discovery.
Self-attribution is the most reliable source of AEO-influenced lead data. Add a “How did you hear about us?” field to your forms with AI tools as an option, broken out by name. Buyers who found you through ChatGPT or Perplexity will tell you, and that first-party data is more accurate than anything analytics can infer.
Cookie attribution covers the rest. Analytics will still record AI-influenced leads, but they typically land under direct or organic. Tracking alongside each other helps triangulate the real volume of AEO-driven enquiries.
This can be seen in recent research by n8n that leads self-attributed – “Where do you hear about us” are 10x that of leads attributed to AI tools via Google Analytics. source
Volume is only half the story. Buyers who arrive via AI tools have usually done meaningful research before reaching out, which tends to show up in two places: a higher close rate and a shorter sales cycle, since the qualification work has already happened inside the AI conversation.
Deal size and customer fit round it out. If AEO is working, you should see not just more leads but better ones: closer to your ideal customer profile, faster to close, and worth more once they do.
AEO is often compared to SEO, since both focus on search visibility. But the value they deliver is different. Strong AEO gets you discovered by high-intent buyers, and it acts as a recommendation. Would you rather send someone to a generic landing page telling them why you’re the best, or have an AI tool recommend your business based on their specific needs?
AEO also goes beyond discovery. Buyers who already know the players use AI tools to compare them and decide which is the best fit. Investing in AEO drives leads while supporting sales and marketing throughout the entire buyer journey.
Visibility Wins builds AEO measurement frameworks alongside the strategy itself, so that progress is trackable from the outset rather than something to figure out after the fact. That means establishing baseline visibility testing for the queries that drive new business, setting up lead source tracking, and defining the qualitative and quantitative indicators that will demonstrate progress over time.
The goal is to give the business a clear, honest picture of where it stands and how that changes, without overpromising what is attributable or obscuring what is genuinely difficult to measure.
Be The Brand AI Recommends
See where your brand appears in AI search, where your competitors are winning and what it takes to become the answer AI recommends
Senior Growth Executive, Visibility Wins
Adam Grant leads the Answer Engine Optimisation (AEO) team at Visibility Wins. Adam specialises in helping B2B IT companies become discoverable across Google and LLMs. His work focuses on building the content and authority signals that get brands recommended as trusted answers.