AI answers
When someone asks an AI for a recommendation in your category, does your brand appear, in what position, on how many platforms? Most companies have no idea.
300+ AI models. 5M+ URLs calibrated. 7 reader segments. One index that tells you where to act.
THE SHIFT
Buyers no longer start with a search box. They ask an AI, scan a feed, skim a page, and decide in seconds. Across all three, your brand is either captured attention, or it was invisible. Almost no one measures which.
When someone asks an AI for a recommendation in your category, does your brand appear, in what position, on how many platforms? Most companies have no idea.
On your own pages and earned media, attention is not clicks or pageviews. It is whether people actually read, how deep they go, and where they drop off.
In the feed, attention is fragmented and interrupted. Reach means nothing if no one stops, reads, or stays.
THE PLATFORM
CAX
Behavioral Biometrics Engine® measuring real human attention on content. 5M+ URLs calibrated over 9 years. Attention decay models.
Explore CAX →BAX
Measures brand visibility in ChatGPT, Perplexity, Gemini and 300+ AI model variants. Citation depth, sentiment, share-of-voice in LLM responses.
Explore BAX →VAX
Measures real attention on video content. Attention retention curves, scroll depth per format, dwell time per post. Publisher dashboard for portfolio management, agency tools for format intelligence and creator ranking.
Explore VAX →AAX
Independent audit of your attention measurement stack across five layers. Documents your stack against the IAB/MRC Attention Measurement Guidelines. For banking, pharma, and regulated industries.
Explore AAX →THE ENGINE
Most tools count mentions. BAX measures cognition. The Behavioral Biometrics Engine® classifies how people actually read, how deep they go, and whether they were paying attention when your brand appeared.
Every user is classified into one of seven behavioral segments: Deep Reader, Active Explorer, Targeted Scanner, Flow Scroller, Headline Skimmer, Distracted Browser, Content Binger. A Deep Reader who reaches your brand has twice the recall probability of a Flow Scroller. BAX weights every metric accordingly.
A brand mentioned first in an AI response reaches 77% of users. Mentioned last, 18%. The same pattern holds on web pages. BAX applies a decay model calibrated on millions of real reading sessions to quantify this across every surface.
Every BAX Index result is replicable. Open formulas, published weights, raw data stored in EU infrastructure for 365 days. Open methodology.
BAX INDEX
BAX Index is a composite index from 0 to 100. Four components, explicit weights, no black box. It measures the quality of cognitive attention your brand actually receives, not just whether you were mentioned.
How many users actually process your brand mention, adjusted for position in the response.
Positive recommendation versus neutral mention versus negative framing.
Depth of attention based on reader segment, time in view, and position centrality.
Factual correctness of what AI says about your brand. Every hallucination costs you points.
HOW IT WORKS
BAX measures what AI says about your brand and what AI crawlers can actually read on your site. Then it tells you what to change, hands your team every change ready to ship, and proves on day 90 which changes worked.
We run your approved question set across the major AI platforms before a single page is touched. The questions follow how people decide: awareness, research, comparison, recommendation. You see where your brand appears, in what position, which sources AI cites, and what it gets wrong.
We test every page in your sitemap, no sampling, in the raw version most AI crawlers read. Crawler access, structured data, expert signals, empty pages, broken links, and whether your content leads anywhere. Findings are ranked by their impact on AI visibility, not by how easy they are to fix.
Not a to-do list. A 90-day playbook split by role, for editorial and for SEO and IT. Every task comes with the URL, the current state, paste-ready text or code, where it goes in your CMS, and an acceptance test.
We re-run the same questions and the same site tests. Changed pages are compared with a control group left untouched, so the effect of your work is separated from changes in the AI models themselves. We report change against baseline. We do not promise traffic.
WHO IT'S FOR
Your brand competes for attention in AI answers, on the web, and in social feeds. BAX shows where you're winning, where you're invisible, and what to fix. FMCG, pharma, financial services, automotive, technology, retail.
You sell attention, but you measure pageviews. BAX shows which content actually captures reader attention, how deep they read, where they drop off, and which sources deliver engaged audiences.
One platform, every client brand, every surface. The audit, the score, the competitor benchmark, and the report that wins the pitch.
BAX — COMMON QUESTIONS
The BAX Index is a composite attention metric scaled 0-100 that measures real human cognitive engagement across AI responses, web content, and video. It combines decay-weighted reach (45%), sentiment (25%), exposure quality (20%), and accuracy (10%) into one comparable number across all channels.
BAX monitors brand visibility across 300+ AI model variants on 6 major platforms including ChatGPT, Perplexity, and Gemini. It tracks citation depth, position, sentiment, share-of-voice, and narrative accuracy in LLM outputs.
The Behavioral Biometrics Engine (BBE) is a proprietary technology that measures real human cognitive attention through behavioral signals captured in standard web browsers — scroll patterns, dwell time, cursor dynamics, and reading depth. It operates at census scale, calibrated on 5M+ URLs over 9 years.
Viewability measures whether an ad had the opportunity to be seen. BAX measures whether a human actually processed the content — scroll depth, active reading time, cognitive engagement state, and attention decay. BAX counts only CLEAN sessions after bot filtration.
BAX classifies every user session into one of seven cognitive segments: Deep Reader, Active Explorer, Targeted Scanner, Flow Scroller, Headline Skimmer, Distracted Browser, and Content Binger. A Deep Reader who encounters your brand has twice the recall probability of a Flow Scroller.
One platform. Every environment where attention matters. Request access and run your first audit.