Home
About OpenLens

A research team studying how AI forms recommendations built a way to track them.

We started OpenLens when we wanted to test our ideas against reality — and found no tools that matched our understanding of how these systems work.

Why This Exists

Stop treating AI search like traditional search.

It is not. Traditional search is an index. You could browse through it, submit a site, and reliably show up. AI search is not browsable. You cannot submit your website to it. And yet, within months, a majority of your potential customers will be using it to choose what they buy.

What you see with traditional search
  1. 1.Buyer types query into Google
  2. 2.Clicks through to your website
  3. 3.Analytics records the visit
  4. 4.Lead traced to source
You see every visit
Attribution is clear
Leads are traceable
What you miss with AI search
  1. 1.Buyer asks ChatGPT or Claude
  2. 2.AI generates a recommendation
  3. 3.Buyer acts on the answer
  4. 4.You see none of it
No clickstream
No attribution
Evaluation invisible to you

Your existing tools were built for a system of pages and links. AI search generates recommendations and answers. You need new tools to see what these models are saying about you. OpenLens was built for this.

Under the Hood

We open the apps. We run your prompt. We read what they say.

You are not checking a position on a page. When someone uses ChatGPT or Perplexity, they do not see ten blue links. They see an answer. You may be in it, or you may not. We monitor these answers: whether your brand appears, how it is described, and which other brands appear alongside it.

1

Query across platforms

We send your prompts to all five AI platforms via their official APIs. For platforms where the app experience differs from the API — ChatGPT, Perplexity, and Google AI — we also run automated browser sessions against the live app. This captures what users actually see, not just what the API returns.

2

Dual-validation extraction

Fast fuzzy matching detects your brand name in each response. An AI parser then extracts structured context: sentiment, position, attributes, and citations. Each result is cross-validated to eliminate false positives.

3

Daily snapshot aggregation

Results are rolled into daily snapshots per brand, keyword, and platform. This creates a reliable time series. When your content changes, you can see exactly which models responded, and how.

The Team
CaltechGeorgia TechUniversity of Toronto

“The AI visibility space is filling up with tools built by people who noticed the trend and moved fast. We took a different route.”

The aibread.com team is composed of AI researchers from Caltech, Georgia Tech, and the University of Toronto. We study how language models form recommendations, and we publish our work in the major AI conferences. When these models became search engines, we tested every monitoring tool on the market. None were built by people who understood how these systems actually work. They applied old SEO thinking to a new kind of technology. So we started over and built something better.

At some point we stopped trying to ask better test questions and started wanting to know:

What would you do with this?

What would make you change your strategy?

What would you need to see to recommend this channel to your clients?

We built OpenLens to find these answers. Since AI visibility is a new space, we are giving OpenLens to the agencies who are figuring it out. There is no established playbook for this channel, so you get to tell us what you need. Use it for your clients. Tell us what is missing. We will build it. You get first-mover access and a tool that fits your specific problems, and we learn what this space requires.

Start tracking your AI visibility.

See how every major AI platform talks about your brand. Set up in under two minutes, no credit card required.