Airuz

01Overview

People increasingly ask an assistant instead of searching. That moves the question from where a brand ranks to whether it gets mentioned at all — and nothing in a conventional analytics stack answers it. Airuz does.

02Challenge

An answer engine has no results page to scrape and no ranking to read. The only honest way to know whether a brand is recommended is to ask the engines the questions its customers would actually ask, and then read what comes back.

03Approach

I built it to work on the first visit: you type a domain and get an answer in under a minute, with no account. Earning a sign-up before showing any value is how a tool like this dies on arrival.

04Solution

Airuz reads the site to infer the brand, its category and its market, generates real buying prompts from that, runs them against ChatGPT, Gemini and Perplexity, and reports how often the brand appears, in what position and alongside which competitors. I built and run the product end to end, in Spanish and English.

05What was built

• Brand, category and market inferred from the site itself
• Buying prompts generated per category, not a fixed question list
• Three answer engines queried live and compared side by side
• Mentions, position and competitors reported per run
• First analysis free and account-free; history behind a free account
• Bilingual product, Spanish and English

06Technology

• Next.js
• React
• TypeScript
• OpenAI
• Gemini
• Perplexity
• Cloudflare