Elementera Lenz

By Dr. Mehrdad Mehrbod, Co-Founder and Chief Product Officer. Reviewed by Soroush Mehri, Co-Founder and Chief AI Officer. Last updated October 6, 2026.

Lenz asks ChatGPT and Gemini the questions a buyer in your category would ask, without ever using your name, and counts how often each engine names your company in its answer. A report gives you a score for each engine, the competitors the engines named instead, and, question by question, who was named.

Lenz is Elementera's AI visibility platform, built for small and mid-sized businesses and funded startups: SaaS, fintech and tech companies, and other growing businesses whose buyers now ask AI before they ask a salesperson. Lenz shows you, engine by engine, where you are named, where competitors are named instead, and what to work on first.

What does Lenz measure?

Lenz measures mentions: whether ChatGPT and Gemini name your company when a prospective buyer asks them about your category without naming you. For each engine, your score is the share of that engine's answers that named you. A score of 30% on ChatGPT means ChatGPT named you in 3 of every 10 answers Lenz collected.

For every business an answer names, Lenz also records its position in the answer (first, second, third), whether it was recommended as a top pick, listed as one option among several, or mentioned in passing, the tone (positive, neutral or negative), and a short quote around the name, so any count can be checked against the engine's own words.

Lenz counts your name in the answer itself, the moment an engine puts you in front of a buyer.

How does Lenz measure it?

Lenz writes blind buyer questions from your own website, asks ChatGPT and Gemini, has a separate model read each answer, and scores each engine on its own. A run has six steps.

1. It reads your website. From your homepage, Lenz proposes your industry, your products and services, and a short description of your business model, and you can correct all three. To write questions it then reads your homepage and up to four more pages, chosen because their addresses suggest FAQ, services, products, pricing or about, and takes up to 3,000 characters from each.
2. It writes the questions. A language model writes buyer questions from that material, from early research ("how do small teams stop doing payroll in spreadsheets?") to ready to buy ("best payroll tool for a seed-stage SaaS company"). The model is told your company name so it can understand the site, and it is instructed not to put your name, domain or initials in any question.
3. You review the list. Before anything is asked, you can delete, edit or add questions, or upload your own list. The list you approve is exactly what runs.
4. It asks ChatGPT and Gemini, without your name. By default, Lenz collects each answer through a third-party data service that, by its own description, gathers answers from the ChatGPT and Gemini web interfaces rather than from a developer API. Only the question is sent, with the location set to the country you choose (the United States if none is set) and the language set to English. ChatGPT is asked to use web search for every question, so answers reflect what the engine finds today. A setting can route the same questions through the engines' developer APIs instead.
5. A separate model reads each answer. An OpenAI model lists every business the answer offers as an option, in order, with the details above. It is not told which business we care about, so it cannot lean towards finding you. A business cited only as evidence ("according to a review site...") is left out, because it is not offered as a choice.
6. It matches and scores. Lenz matches the names it found to your company by fixed rules that ignore "Inc.", "LLC", capitals and punctuation, then scores each engine separately. The report records how many answers each engine returned, so every percentage can be read as a count. The overall figure is the average of the ChatGPT and Gemini scores.

What does a Lenz report look like?

A Lenz report shows an overall visibility figure, your score against the most-named competitors on each engine, and sample questions with the businesses each engine named. You can download a PDF summary and a spreadsheet (CSV) with one row per question and engine: the question, the engine, whether you were named, and which other businesses were named.

A sample report, with an invented company

Everything in this sample is invented: the company (Example Payroll Co., at example.com), the four competitors (Competitor Alpha, Bravo, Charlie and Delta), the questions and every number. No client data is used, and none of these names refers to a real business. To make the arithmetic match what Lenz does, we produced the figures by feeding invented answers through Lenz's own scoring code. The sample has ten questions.

The headline.

Your brand against competitors. By default, the report shows the three competitors with the highest overall figure.

Competitor Delta scored 25.56% overall. The report lists three competitors, so Delta is left off, yet it still counts in the rank, which is why the sample company is 5th and not 4th.

Who the engines named, question by question. The report shows this for the first five questions; the CSV covers all of them.

How do you get the most from a Lenz report?

Read each engine on its own, turn percentages back into counts, and act on the questions where competitors are named instead of you.

1. Read each engine on its own. ChatGPT and Gemini draw on largely different parts of the web, so each has its own score and its own competitor list.
2. Turn the percentage back into a count. 30% of ten answers is three answers, which makes progress easy to track.
3. Look at how you were named. The breakdown separates a top recommendation from a passing mention, so you can see where you are already trusted.
4. List who was named where you were not, and how they were named. That list is your shortlist of what to work on first.
5. Add questions in your buyers' words. Take three from sales calls, demo requests and support tickets, then save the report and re-run the same list to track change week to week.

What is Origio, and how does it relate to Lenz?

Origio is Elementera's emerging original-content workflow. It interviews you and your team to capture real experience, examples and point of view, then shapes that into original material. It is developed and tested alongside Lenz, and it does not score or rate content.

The reason it starts with an interview is simple: a language model cannot supply facts it does not have. When we asked two OpenAI writer models to write expert copy on 38 questions whose answers exist only in our own unpublished measurements, none of 71 answers got the fact right. Original information has to come from people or from measurement, so Origio starts from what your business knows.

Frequently asked questions

Is Lenz free? A new Lenz account includes one free analysis. After that, starting a new analysis or re-running one takes you to book a call with us.

How long does a Lenz analysis take? It runs while you wait, and the screen shows progress question by question.

Which engines does Lenz check? ChatGPT and Gemini, the AI assistants buyers use most. Claude, Perplexity and Grok also appear in the report table.

Can Lenz tell me why I am not named? Lenz shows where you were missed, which competitors were named instead and how they were named, which is the clearest starting point for fixing it. Our managed AEO services take it from there.

Work with us

There are two ways to work with us. Teams that want the work done for them can use our managed AEO services: Book a call. Teams that want to see where they stand first can start with Lenz: Get your free AI visibility report.