Numbers you can actually trust — humans, bots, and AI, told apart
By most estimates more than half of web traffic is automated, inflating your sessions, skewing your conversion rates, and hiding AI referrals in “Direct.” Some of that “Direct” is AI sending you buyers who tend to convert well above your organic traffic. Lume gives you a clean human vs. bot vs. AI-agent split, pulls those AI-referred visitors out of the dark, and hands each number a next step: which channel to double down on, which surface is converting.
Sound familiar?
- My traffic is up but conversions are flat. Turns out a lot of it is bots.
- I can’t trust my dashboards; bots inflate sessions and wreck bounce and conversion rates.
- Is AI referring visitors? I can’t tell. It’s all lumped into “Direct.”
- I’m making budget decisions on numbers I don’t fully believe.
What Lume does for you
A human / bot / AI split you can trust
Lume classifies every request as human, bot, or AI agent, server-side, so it catches the bots that don’t run JavaScript and slip past client-side analytics.
Strip the bot noise
See your real human numbers, so bounce rate, session counts, and conversion rates reflect people, not crawlers padding your dashboards.
Measure AI as a channel
See the human visitors ChatGPT, Claude, Gemini, and Perplexity send you, detected by referrer host and utm_source even when the referrer is stripped, so you can tell which AI surface is actually converting and put budget there.
Prove what drives real visits
Attribute AI-driven human traffic instead of losing it to “Direct,” then export it back into the analytics and warehouse you already use, so AI shows up as a real channel in the funnel and dashboards your team already lives in.
Fits your stack
Report traffic from WordPress, your CDN, or the edge: low-cost, always-on instrumentation that runs alongside the analytics you already have and exports back into them, so it’s one more trustworthy number, not one more login.
Marketing FAQ
Does this replace Google Analytics?
No, it complements it. Lume runs server-side and catches bot and AI-agent traffic that client-side analytics miss, then gives you a clean human vs. bot vs. AI split to read alongside your existing tools.
Why is my “direct” traffic actually AI?
AI apps often strip the referrer, so those visits fall into Direct. Lume detects them by referrer host and utm_source and attributes them to the assistant that sent them.
How do I separate real visitors from bots?
Every request is classified as human, bot, or AI agent, and the major operators are verified by signature, reverse DNS, or published IP ranges, so your human count reflects people, not spoofed crawlers.
Can I prove AI drives conversions?
Lume attributes AI-referred visits so you can follow them into your own funnel and see whether AI answers bring buyers, not just clicks.
Get numbers you can actually trust.
Free plan, one ingest token. Clean out the bot noise and see the AI traffic you’re already getting.
Start for free →