July 7, 2025·News

Measuring Traffic from LLMs

News in digital marketing

There’s a wave of new tech flooding the marketing scene, and fresh tools are popping up to help companies track brand visibility in AI platforms. The share of users visiting websites from LLM links is growing fast, but it’s still tough to pinpoint exactly where this traffic comes from or how sites rank in ChatGPT, Claude, Grok, Gemini, Copilot, and the rest. That’s about to change.

Marketers are starting to realise that LLM visibility is becoming critical, but there are still major gaps creating pain points:

  • Lack of brand mention tracking across platforms and prompts, making it hard to know which keywords and pages drive AI exposure or how your share of voice stacks up against competitors.
  • No clear way to model traffic coming from LLMs to your site. As organic search keeps declining, traditional analytics can’t see where new growth is actually coming from.
  • Limited actionable insights, with few options to auto-apply recommendations to ad copy, images, or even on-page SEO like titles and headlines.

Here are the companies moving first to solve these problems:

Adobe

Announced in June 2025, Adobe LLM Optimizer lets brands track where and how their name pops up in AI-generated results. The big differentiator? It taps into Adobe’s full marketing suite to not just report mentions, but also recommend changes to your website, ads, and content—and then deploy them directly through A/B testing, social publishing, and creative updates (assuming you use those services).

AHrefs

Brand Radar tracks brand mentions across LLM platforms, showing where your brand appears alongside keywords and ranked pages. AHrefs, Moz, and SEMrush are still SEO staples, so it’ll be interesting to see how fast SEO specialists add this extension to their toolkit.

HubSpot

No plans yet for LLM mention tracking or visibility optimisation. Their latest move to position as AI-friendly is Breeze, HubSpot’s AI suite with features like generative content, automated sales outreach, customer support assistants, and CRM enrichment. I wouldn’t be surprised if LLM tracking gets rolled in there soon.

The DIY Alternative

Most of us are in this territory. Talking to a few people in the field, the concern for Organic drop is growing, and the results of manually prompting ChatGPT is now being included in marketing reports. But thanks to AI coding capabilities, spinning up a custom solution is now within reach for non-developers too. As I wrote this post, I built a quick proof of concept to check brand mentions for airlines across two prompts:

import os
from openai import OpenAI

client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))

prompts = [
    "Which airline has the best long-haul business class experience?",
    "What is the top-rated airline for long-haul flights globally?"
]

airlines = ["Emirates", "Qatar Airways", "Singapore Airlines"]
mentions = {airline: 0 for airline in airlines}

for prompt in prompts:
    response = client.chat.completions.create(
        model="gpt-4o",
        messages=[
            {"role": "system", "content": "You are a helpful assistant."},
            {"role": "user", "content": prompt}
        ]
    )
    answer = response.choices[0].message.content
    for airline in airlines:
        if airline in answer:
            mentions[airline] += 1

print(mentions)

Example output:

{'Emirates': 2, 'Qatar Airways': 1, 'Singapore Airlines': 1}

Not bad for five minutes of work. You can extend this to track more prompts, calculate share of voice percentages, or integrate it with dashboards for weekly competitive benchmarking. Or tailor it to your niche.

And if you’ve never coded before? Just ask your AI tool of choice to walk you through making this script work on your machine. It's easier than you think, but it basically boils down to this:

  1. Get API keys for the LLM you want to track (e.g. OpenAI)
  2. Install an IDE to work with code, paste the script, and run it (Visual Studio Code usually does the job).
  3. Screenshot any errors back to the LLM for debugging.
  4. Keep pasting the new code in until it works.

That’s all there is to it. Honestly, after a few days of trial and error, you’ll probably get something close enough to reality.

What's Coming

This is a massive opportunity for companies to offer customer-centric solutions with robust LLM tracking and actionable recommendations. For marketers, experimenting with simple scripts like this builds habits of measuring and implementing regularly, just like checking your Google rankings, so you’re ready when more polished tools roll out. But one thing is for certain: the quality, freshness, and uniqueness of the content is going to be twice as important as before. Simply because now there's two algorithms to convince: that of the Search Engine, and that of the LLM searching it.

As always, if you want to learn more about what this means for your business and how to optimise your marketing, reach out: Contact Us

Join our newsletter