An AEO agency in India. Here is exactly what we do.

Founders and marketing leads at Indian companies whose buyers ask an AI assistant before they buy: D2C brands, SaaS, clinics, edtech, marketplaces. This site assumes you have never heard of AEO. Most people we talk to have not. By the time they call us, the AI assistants are already recommending someone else.

We have measured 20 Indian brands across 18 categories and one full category, payroll and HRMS, on four engines. The numbers are published, winners and zero-named brands alike, on the research page and the case study.

1. Audit. We run the real questions your buyers ask through ChatGPT, Perplexity, Gemini and AI Overviews. We log who gets named, in what order, and which pages each engine read to decide.
2. Fix. We open the site to every engine's crawler. We mark up what is already published with the schema each engine reads, and add an llms.txt. Then we hand your team a page-by-page brief for what is missing. Your writers write; we check it.
3. Publish. On your site, in your voice, checked by you before anything goes live.
4. Monitor. We re-run the questions every week. Once a month you get one report, in plain numbers: named or not, per engine, against the competitors that beat you.

• A written baseline: your share of answers, per engine, per question, with the sources each engine read.
• Crawler access fixed for GPTBot, PerplexityBot, Google-Extended and ClaudeBot, including the CDN and firewall settings that silently block them.
• Schema on pricing, comparison and FAQ pages: Organization, Product, Offer in rupees, FAQ.
• An llms.txt, a clean sitemap, and one consistent set of company facts across the site.
• A page-level brief for every buyer question you are losing: which page answers it, what it must state, in what order.
• An audit of the third-party record, G2, Capterra, SoftwareSuggest, Reddit and your category's roundups, for missing, wrong or rupee-less entries.
• One report a month. It starts with what the work was worth. Then how your share of answers moved on each engine since we started. Then what we shipped, with quotes from the answers before and after. Then no more than three things to do next month. If nothing moved, it says nothing moved.

A mention is your brand appearing as a whole word in an answer. Your share is the number of answers that name you, divided by the number of answers. We score it per engine. We also score questions that contain your name apart from questions that do not. The unbranded number is the one that matters, and it is the one we report on.

For clients, we ask the same buyer questions every week, through each engine's live search. Same questions every time, so the numbers can be compared. Answers vary between runs, and the consumer apps add personalisation the API does not capture. So we report trends, not single readings. An engine we have not measured is reported as not yet measured, never assumed.

We do not promise rankings, and we do not guarantee an outcome, because nobody controls what an engine writes. We promise to measure your citation share honestly. We promise to do the work that moves it. And we promise to tell you in plain numbers what moved and what did not.

We do not write your pages. Your team knows the product. We tell them exactly what each page has to say for an engine to quote it. Then we check it before it goes live.

We do not publish a rate card. The work differs with how many buyer questions you are losing, and on how many engines. The intro call costs nothing. Fifteen minutes, we run ten of your buyers' questions live and show you who gets named. You leave with the baseline whether or not we work together.