At Big 4 firms, AI is starting to reshape the tax profession
Mumbai, Jun 1, 2026
Synopsis
Major accounting firms in India are embracing artificial intelligence for tax services. AI is automating routine tasks, research, and drafting, allowing consultants to focus on complex interpretations and client strategy. This shift is significantly increasing efficiency and changing traditional consulting models. Firms are seeing substantial growth in AI query volumes, indicating widespread adoption and a new era for tax advisory.
As the Indian arms of the Big Four race to deploy artificial intelligence, tax is emerging as the first major battleground, with firms redesigning workflows in a profession where large teams of consultants have long helped clients navigate the complex regulatory maze of GST, direct taxes, customs duties, transfer pricing and litigation.
The level of adoption is high across companies as AI is increasingly being used by tax consultants to do routine, preparatory and research-intensive work.
At PwC India, around 1,500 employees generate more than 10,000 AI queries daily. At Deloitte India, up to half of its 5,000-member tax team uses proprietary AI platforms in daily work, while at EY India, more than 75% of its 6,500 tax professionals regularly use AI tools in their day-to-day work.
“We have already crossed over 1.3 million queries on the system this year,” said Bhavin Shah, partner, Price Waterhouse & Co LLP. “The scale at which tax professionals are using AI internally is significant, and we are seeing query volumes grow nearly 30% every 45 days.”
In the last 12 months, all the top firms have been working feverishly on building tightly controlled AI ecosystems trained on years of proprietary tax knowledge, past opinions and top legal databases, while trying to weave them in daily workflows of tax consultants on both compliance and advisory side of business.
“Generative AI is probabilistic, while tax is fundamentally deterministic,” said Sameer Gupta, tax leader, EY India. “Clients ultimately need a precise answer on tax liability, merger structures or litigation positions. We have to bridge that gap.”
Experts say few markets offer a better test case for AI in tax than India. It is a market where tax has large transaction volumes, frequent regulatory updates and lots of litigation, making it a function that involves not only large amounts of data but is also heavily dependent on interpretation.
“In India, one transaction can simultaneously trigger GST, customs, transfer pricing and capital gains implications,” Gupta said. “You also need to understand how courts and tribunals may interpret a position. The challenge is not just understanding the facts, but reframing and refining them through the right questions.”
With AI, research that once took several days can now be completed in minutes or hours. Tax associates can generate more comprehensive first drafts, while partners spend more time studying interpretations, searching for identifying risks and applying judgment.
“AI is not replacing expertise; it is amplifying it. The combination of deep domain knowledge and AI-driven intelligence is enabling our consultants to deliver sharper insights, faster turnaround times and more tailored outcomes for clients,” said Gokul Chaudhri, president, Tax, Deloitte South Asia.
With AI deployment, the traditional consulting model of scaling revenue by adding junior staff for research, drafting and compliance work, a structure that has long been followed in the tax business as well, is starting to shift as AI automates large parts of that pyramid.
“The expectation now is to deliver significantly higher revenue growth without proportionate increases in headcount,” Shah said. “The pyramid structure is already beginning to change.”
“Digital workers and AI agents could eventually become an integral part of how firms think about workforce capacity itself,” Gupta added.
Tax leaders say the next phase of the AI race will revolve around agentic AI, where multiple AI agents work together simultaneously across different tasks and workflows, much like teams of professionals collaborating on a project.
For instance, one AI agent could undertake legal research, another could summarise findings, a third could draft presentations or responses, while a supervisory AI system coordinates the entire process.
“That orchestration layer is where the future is heading,” Shah said.
Since tax in India is a high-stakes domain where mistakes lead to litigation, penalties and reputational damage, the firms are wary of hallucinations, unverified information and data privacy issues while using AI and are putting in necessary guardrails. “The biggest risk lies in how the problem is framed at the beginning,” Gupta said. “If contextual inputs are incomplete or incorrect, the AI engine may start from the wrong premise.”
The transformation is particularly being noticed in compliance-heavy areas such as GST.
Gupta said EY is already using OCR and AI tools to read invoices, classify transactions and determine GST input-credit eligibility. The tools continuously learn from corrections made by professionals, improving classification accuracy over time.
“We are increasingly building AI-driven tax agents directly within client environments,” Gupta said. “The idea is to identify misclassifications and compliance risks at the source itself rather than after transactions are completed.”
Tax leaders have also recorded that the sophistication of AI adoption differs across tax domains.
“Corporate tax and litigation-heavy direct tax areas are seeing faster traction because of the large volume of judgments and precedents involved. Transfer pricing and international taxation remain more fact-intensive and contextual, which means human interpretation continues to play a bigger role there,” said Gupta.
Inside the firms, the AI shift is also reshaping training and talent management.
PwC’s Shah said it conducted in-person training sessions across multiple offices and built self-learning modules to accelerate AI adoption internally.
“Leadership teams were also given visibility into usage patterns to push adoption across practice groups,” he said.
At EY, Gupta said the firm embedded technology and tax specialists directly within the same teams rather than outsourcing development externally.
“Our tax experts define the business problems while in-house AI specialists build the solutions alongside them,” Gupta said. “That integration has been critical.”
What many tax leaders are now grappling with is how billing models will evolve when clients begin to question paying traditional fees for work that is increasingly being done by AI.
To prepare for that shift, the firms are building AI platforms that can be sold directly to corporate tax departments on a subscription basis.
PwC said hundreds of clients have already tested its platform, with more than 100 moving to subscription-based models.
Each firm is pursuing a different technology architecture in its race to scale AI in tax in India.
At Deloitte India, Tax Sphere, a unified digital tax platform, is the backbone of its AI strategy in tax, bringing together around 10 technology platforms, including Tax Pragya, into a single ecosystem that works across compliance, litigation, analytics, advisory and workflow automation.
“The integrated ecosystem is helping organisations move from fragmented tax processes to connected, data-driven decision-making models. Technology is enabling compliance to become more scalable and efficient, while making tax advisory services significantly more data-rich, predictive and robust,” said Deloitte’s Chaudhri.
Gupta said EY, the market leader in tax by a margin, began investing in tax technology nearly a decade ago, long before generative AI became mainstream.
The firm has put together a dedicated tax technology team embedded within its tax practice rather than outsourcing development.
Tax specialists and technologists work together in a “two-by-two matrix,” jointly responsible for creating tools that solve real client problems.
“We may be the only firm where both the technical and AI teams sit within the tax practice itself,” he said. “That makes the development strategic rather than tactical.”
"We have already crossed over 1.3 million queries on the system this year," said Bhavin Shah, partner, Price Waterhouse & Co. "The scale at which tax professionals are using AI internally is significant, and we are seeing query volumes grow nearly 30% every 45 days."
In the last 12 months, all the top firms have been working feverishly on building tightly controlled AI ecosystems trained on years of proprietary tax knowledge, past opinions and top legal databases, while trying to weave them in daily workflows of tax consultants on both compliance and advisory side of business.
"Generative AI is probabilistic, while tax is fundamentally deterministic," said Sameer Gupta, tax leader, EY India. "Clients ultimately need a precise answer on tax liability, merger structures or litigation positions. We have to bridge that gap."
Experts say few markets offer a better test case for AI in tax than India. It is a market where tax has large transaction volumes, frequent regulatory updates and lots of litigation, making it a function that involves not only large amounts of data but is also heavily dependent on interpretation.
"In India, one transaction can simultaneously trigger GST, customs, transfer pricing and capital gains implications," Gupta said. "You also need to understand how courts and tribunals may interpret a position. The challenge is not just understanding the facts, but reframing and refining them through the right questions." With AI, research that once took several days can now be completed in minutes or hours. Tax associates can generate more comprehensive first drafts, while partners spend more time studying interpretations, searching for identifying risks and applying judgment.
"AI is not replacing expertise; it is amplifying it. The combination of deep domain knowledge and AI-driven intelligence is enabling our consultants to deliver sharper insights, faster turnaround times and more tailored outcomes for clients," said Gokul Chaudhri, president, Tax, Deloitte South Asia.
With AI deployment, the traditional consulting model of scaling revenue by adding junior staff for research, drafting and compliance work, a structure that has long been followed in the tax business as well, is starting to shift as AI automates large parts of that pyramid.
"The expectation now is to deliver significantly higher revenue growth without proportionate increases in headcount," Shah said. "The pyramid structure is already beginning to change."
[The Economic Times]
