I scanned 16 Indian CA and accounting firms for AI visibility

On 1 August 2026 I scanned 16 Indian chartered accountant and accounting-services firm homepages with my own AI-readability engine, the same one behind the free Report Card tool. It grades a page 0 to 100 on the technical signals that let AI answer engines like ChatGPT, Perplexity and Google AI Overviews read, quote and cite a business: structured data, FAQ schema, Open Graph tags, page title and a single clear H1. Accounting turns out to be a professional-services vertical that reads well to AI, unlike law: the median score is 87, against 59.5 when I scanned Indian law firms. The reason is not the traditional CA practice. It is the digital-first platforms that turned company registration and tax filing into SaaS-style products, and pulled the whole vertical's median up with them.

The result: a median of 87, and a clear split by firm type

Across the 16 sites, the median score was 87 out of 100 and the mean was 82.7. Eleven firms scored an A (85 or above), three scored a B, and two scored a C. Nobody scored a D or an F. The range ran from 58 to 92, so even the weakest site in the sample cleared halfway.

That is a markedly stronger result than the other professional-services vertical I have scanned. Indian law firms came back with a median of just 59.5. Accounting sites, on the whole, are simply easier for an AI answer engine to read.

The median hides a real split. Traditional network firms scored a median of 89.5 across 4 sites. Digital-first compliance platforms scored a median of 87 across 5 sites. Regional and independent CA practices trailed at a median of 72 across 7 sites, the weaker tail that pulls the vertical average down from where the leaders sit.

Structured data is the dividing line

Eleven of the 16 firms publish JSON-LD structured data on their homepage. Those 11 sites have a median score of 92. The five that don't have a median of 72. That is a 20 point gap tied to a single signal.

Structured data is how a page tells an AI engine what it is, in a format the engine doesn't have to guess at: this is an accounting firm, here is the service, here is the location. Without it, an AI has to infer that from unstructured prose, and inference is where firms get left out of the answer.

The split lines up with firm type almost exactly. The digital-first platforms and the larger network firms are the ones publishing JSON-LD. The regional and independent practices, the weaker tail at a median of 72, are disproportionately the ones without it.

The FAQ-schema gap: 1 of 16

Only one of the 16 firms has FAQ schema on its homepage. Fifteen do not, in a vertical built almost entirely around answering the same handful of questions on repeat: how to file GST, how to register a company, what a statutory audit costs, what to do about a tax notice.

FAQ schema is the exact format for that. It marks up a question and its answer in a way an AI engine can lift directly and cite, instead of a person having to skim a paragraph of brochure copy to find the number or the step. For a vertical this question-heavy, it is close to a free citation.

It is also the cheapest fix in this entire study. A firm doesn't need to rebuild its site or rewrite its positioning to add it. It needs to take the questions its own team already answers every day on calls and in emails, and mark them up.

The smaller gaps in the sample:

• FAQ schema present on only 1 of the 16 firms scanned.
• The 5 firms without structured data scored between 58 and 72.
• Open Graph tags present on 14 of 16, still 2 sites missing.
• 13 of 16 have a single clear H1, 3 don't.

What to fix, for the sites still behind

1. Add JSON-LD structured data if your homepage doesn't have it: the single biggest gap, worth a 20 point median swing.
2. Add FAQPage schema for the questions you already answer every day, on GST filing, registration, audit cost, notices. My free schema generator builds the markup for you.
3. Keep one single, clear H1 that states what the firm does and for whom.
4. Add Open Graph tags so the page previews correctly when it's shared or cited.
5. Run your own homepage through the free Report Card tool to see which of these five signals you're missing right now.

Method and limits

I scanned all 16 firm homepages live on 1 August 2026, using the same public engine anyone can run for free. This looked only at the homepage, not the full site, so a firm with strong FAQ or structured data three clicks deep wouldn't show it here. It's a snapshot in time: sites change, and a rescan next month could move individual scores. The sample is anonymous and aggregate by design, and it skews toward larger and more visible firms, the ones easiest to find and scan, so the long tail of small local CA practices not covered here is likely weaker than what's reported.

Frequently asked questions