Skip to content
aelo
September 2026 · Preview Edition

The India AI Visibility Index

Published observations of AI answers about Indian brands. Every result links to selected samples and their settings. These are API observations, not a claim about what every user sees in an assistant app.

0 brands · 0 categories·No published results

No reviewed results have been published for this edition.

This is not a zero-visibility result, and no running scan is implied.

Each percentage describes only the published samples. Zero mentions does not mean a brand is invisible. Intervals express repeat-sample uncertainty, not coverage of all users. Overlapping intervals do not establish a ranking. List position is not a recommendation score.

Methodology

Every number is one query away from the raw scan.

Matching measurement settings
An edition includes only explicitly reviewed samples with a matching prompt, engine, model, region, mode and scoring versions. A published brand needs at least four successful samples. This limited selection is not representative of every buyer question.
Honest data policy
Missing or failed measurements are not zero visibility. Zero percent means no mention in the published samples, not that the brand is invisible everywhere. Only explicitly approved publications are exposed; private workspaces stay private.
Rates, sample counts and uncertainty
Mention rate is mentions divided by successful samples. Each row includes a 95% Wilson interval. These describe repeat-sample uncertainty under a statistical independence assumption, not the experience of all users or proven market dominance.
How brands are picked
Entries are selected for a reviewed publication. Selection is not a claim that a brand leads its category. Inspect each receipt and the scope before drawing conclusions.

Future editions

Want your brand measured?

Apply to be included in the next edition. If your brand fits a category we track, we can discuss a reviewed measurement. Inclusion, timing and favorable results are not guaranteed.

Apply to be measured Read our honest-data manifesto →

AI answers vary. Model behavior, prompt wording, search settings and timing can all change a result. There is no universal ±10-point error bound. Use the actual sample counts, intervals and published responses to judge what each result supports.