Audit and accounting firms in India are using AI in three distinct ways in 2026. The Big 4 are embedding AI agents inside global audit platforms used by their Indian teams. ICAI is putting free and low-cost tools in the hands of members. And mid-sized and smaller CA practices are using general assistants for drafting, summarising and reconciliation, often without written rules. The firms getting real value share one habit: they treat AI as a change to a specific workflow, with a reviewer, not as a tool everyone experiments with.

I spent my early career in Big 4 and now help CA practices and finance teams implement AI, so I see both ends. Here is what is actually happening.

1. The Big 4: AI inside the audit platform

The large networks run global audit platforms that their Indian member firms use, and those platforms now have AI built in.

  • EY announced in April 2026 that it is embedding a multi-agent framework into EY Canvas, which it says covers the daily work of 130,000 assurance professionals across 160,000 audit engagements in more than 150 countries and processes over 1.4 trillion lines of journal entry data a year. EY expects it to support all end-to-end audit activities by 2028, while stressing that human judgement and scepticism remain central.
  • Deloitte has added a network of AI agents to Deloitte Omnia, its global audit platform, available to nearly 85,000 audit and assurance practitioners, and describes its approach as “human-led, AI-powered”.
  • On the tax side in India, EY India says its India Tax Platform includes AI agents for tax research, compliance and litigation, supports more than 4,000 clients and is built by over 650 tax technologists.

The lesson for everyone else is not to copy the technology. It is that even firms with these budgets keep a person accountable for each conclusion and invest heavily in training alongside the tools.

2. ICAI: tools for every member

ICAI has moved faster than most professional bodies. It offers CA GPT free to members, which it says more than 70,000 members were using by November 2024. It reports more than 150 GPT-based tools and over 50,000 members trained in AI. And ICAI VERA is a digital audit suite built around the Standards on Auditing, with engagement letters, materiality calculators, an evidence vault, Tally and GST data tools and report generators for SA 700/705/706, CARO 2020 and IFC reporting. VERA’s own disclaimer is worth repeating: final professional responsibility for audit conclusions lies with the signing CA.

3. Mid-sized and smaller CA practices

This is where most of the profession sits. ICAI’s 2024-25 membership data shows about 1.61 lakh members in practice. In the practices I work with, AI use typically looks like this:

Patterns from my implementation work, not survey data.
UseWhat worksWhat goes wrong
Drafting emails, notice replies, engagement lettersBig time saving on first draftsWrong sections or rates sent without checking
GST, TDS and bank reconciliationsFast exception lists for a reviewerClient data pasted into personal accounts
Audit analytics and ledger scrutinyTesting full populations, unusual-entry flagsOutput not linked back to working papers
Summarising agreements and circularsReviewer knows where to lookSummary treated as the reading

This matches the wider picture. Microsoft and LinkedIn’s 2024 Work Trend Index found 92% of Indian knowledge workers use AI at work, while 54% of Indian leaders worried their organisation lacked a plan to implement it. In CA practices, that gap shows up as juniors using tools quietly and partners not knowing what client data has gone where.

What a smaller firm should copy first

  1. Write the data rule before buying anything. One page: what may go into which tool, under which account. Use the AI controls checklist for CA firms. Where personal data is involved, the Digital Personal Data Protection Act, 2023 expects reasonable security safeguards.
  2. Pick one workflow with volume. Reconciliations and document-heavy review are good candidates. See AI for GST reconciliation and AI for audit working papers.
  3. Measure before and after. Reviewer time per item and items sent back at review. Run it as a time-boxed pilot using my 90-day AI pilot playbook for CA firms.
  4. Keep the signature human. Every big firm deploying agents says the same; your peer reviewer will too.

The real difference between firms

It is not the tool. Big 4 teams, ICAI VERA users and a ten-person practice with a business AI licence can all get value, or all create risk. The difference is whether someone owns the workflow, the data rule and the review. If you want help mapping where AI pays back in your firm, see the AI Opportunity Audit, start with the AI for chartered accountants in India hub, or book a discovery call.

Frequently asked questions

How is AI being used in accounting firms?

Mostly for drafting client communications and notice replies, first-pass GST, TDS and bank reconciliations, ledger scrutiny and audit analytics, and summarising long documents for a reviewer. The largest firms are also embedding AI agents into their global audit platforms.

How are audit firms using AI?

Large networks such as EY and Deloitte have built AI agents into their audit platforms (EY Canvas and Deloitte Omnia) for risk assessment, testing and documentation, while keeping auditors accountable for conclusions. In India, ICAI offers members CA GPT and the VERA audit suite built around the Standards on Auditing.

How can accounting firms use AI safely?

Write a one-page data rule, use firm-controlled business accounts, start with one high-volume workflow, measure reviewer time and errors before and after, and keep a named person responsible for every output that reaches a client or a filing.

Will AI take over Chartered Accountant work in audit firms?

AI is taking over routine testing, data extraction and drafting, but audit opinions, professional scepticism and the signature remain with a qualified CA. Even firms deploying AI agents at scale describe their approach as human-led.

Sources

Disclaimer: this article is general educational commentary from my implementation work. It is not legal, tax, audit, career or data-protection advice. Fees, courses and tool terms change; check the current details at the source before acting.