AI in the Finance Function: How India's CFOs Are Automating FP&A, Audit and Treasury

At a leading Indian private sector bank, the finance team's monthly close process used to take eleven working days. Hundreds of reconciliation steps, journal entries, and consolidation processes had to be completed in sequence, with each step reviewed by a human before the next could begin. As recently as 2022, this was considered industry-standard performance — perhaps even slightly better than average. By the end of 2024, the same close process was taking four days. The reduction was not achieved by adding headcount or working nights. It was achieved by deploying AI-powered reconciliation and journal entry automation that handled the vast majority of routine steps autonomously, flagging only the exceptions that required human judgment.

That example is one of hundreds that Gladwin International has documented through advisory conversations with CFOs at Indian financial services institutions, listed companies, and PE-backed enterprises over the past eighteen months. The deployment of artificial intelligence in the finance function — which was a theoretical discussion two years ago — is now a operational reality for a meaningful and growing proportion of India's corporate sector. And the CFOs who are leading this deployment are discovering that AI changes not just the efficiency of the finance function but its fundamental character: what it does, who it employs, and what it contributes to the organisation's strategic capability.',

The State of AI Adoption in Indian Finance Functions

The adoption of AI in Indian finance functions spans a spectrum from basic robotic process automation (RPA), which has been in use for several years, to the deployment of large language models (LLMs) and generative AI for tasks that require interpretation, synthesis, and generation of novel content. Understanding where organisations sit on this spectrum is important because the governance, capability, and organisational implications differ substantially at each stage.

A survey by the Institute of Chartered Accountants of India (ICAI) conducted in the second half of 2024 found that 64% of large Indian companies (defined as those with annual revenues above ₹500 crore) had deployed at least some form of AI or advanced automation in their finance function. Of these, 71% had implemented RPA for routine transactional processes, 48% had deployed machine learning models for specific analytics applications such as revenue forecasting or accounts receivable aging prediction, and 23% had begun deploying generative AI tools — either commercial platforms like Microsoft Copilot integrated with financial systems, or internally developed LLM applications.

The 23% figure for generative AI adoption is significant. It represents a doubling from approximately 11% in early 2024, suggesting that the adoption curve is steep and that the proportion will be substantially higher by the time this analysis is read. More important than the aggregate figure is the pattern of adoption: generative AI deployment in finance functions is concentrated in organisations with stronger technology infrastructure, more technically oriented finance leadership, and higher proportions of analytically trained finance professionals — suggesting that the technology is amplifying existing organisational capability advantages rather than equalising the field.',

FP&A: The Biggest Transformation

Financial Planning and Analysis is the domain within the finance function where AI is producing the most visible and consequential changes. The traditional FP&A model — in which finance analysts spend the majority of their time gathering, cleaning, and reconciling data from multiple systems before they can begin any actual analysis — is being fundamentally disrupted by AI tools that can perform data aggregation and reconciliation autonomously.

The consequence is a reallocation of human time from data work to analytical work. At companies where AI-powered FP&A has been deployed effectively, the proportion of analyst time spent on data gathering has fallen from 60–70% to 15–20%, with the freed capacity redirected toward scenario modelling, business partner support, and strategic analysis that was previously squeezed out by the demands of the data assembly process.

At Bajaj Finance, one of India's most analytically sophisticated financial services companies, the FP&A function has deployed a combination of machine learning models for loan book performance forecasting and LLM-based tools for narrative financial analysis. The result is that the monthly management pack — a comprehensive financial and operational review document that previously required ten to twelve analysts working for a week — is now largely auto-generated, with human analysts reviewing, contextualising, and adding qualitative judgment to a first draft produced by the AI system. The number of analysts required for the process has not significantly changed; what has changed is what they spend their time doing.',

"The conversation I am having with my FP&A team is not 'will AI replace you?' It is 'what do you want to become when the machine takes over the work that takes your talent for granted?' The analysts who lean into that question — who are excited about doing more analysis and less data work — are going to have remarkable careers. The ones who see AI as a threat to their current job description are right that their current job description is going away, but they are wrong to see it as a threat." — CFO of a Nifty 50 financial services company, speaking at a Gladwin International CFO Leadership Forum, May 2025.

Internal Audit: Risk Intelligence at Scale

Internal audit is another domain where AI is creating capabilities that were previously unavailable. Traditional internal audit operates on a sampling basis — auditors examine a representative sample of transactions to draw conclusions about the quality of the control environment across the population. Sampling is used because manual review of every transaction in a large organisation is simply not feasible. AI removes this constraint.

At several Indian banks and financial institutions, AI-powered continuous auditing systems now analyse every transaction in real time, comparing each to learned patterns of expected behaviour and flagging anomalies for human investigation. The volume of exceptions flagged is large — AI systems are sensitive enough to detect patterns that human sampling would miss — but the quality of the exceptions is also significantly higher, because the AI is identifying genuine statistical outliers rather than simply selecting items for review at random.',

The implications for the Chief Audit Executive and the CFO — who typically has oversight responsibility for internal audit — are significant. Continuous AI-powered audit changes the nature of the audit finding from a retrospective observation about a past period to a real-time signal about an emerging risk. This enables faster intervention, more precise remediation, and a fundamentally more dynamic risk management environment.',

HDFC Bank, Kotak Mahindra Bank, and ICICI Bank have each invested significantly in AI-powered fraud detection and compliance monitoring systems that are effectively extensions of the internal audit function operating in continuous real time. The savings in prevented fraud and regulatory fines — which are difficult to quantify precisely but are understood by finance leadership at these institutions to be material — represent a direct return on the technology investment.

Treasury: From Reactive to Predictive

Treasury management is the third domain of significant AI transformation in the finance function. The traditional treasury function manages liquidity, interest rate risk, foreign exchange risk, and counterparty risk primarily through rule-based frameworks and periodic manual reviews. AI is enabling a shift from this reactive, rules-based model to a predictive, data-driven model that can anticipate funding needs, optimise hedging positions, and manage counterparty exposure with a level of precision and speed that manual processes cannot match.

Indian companies with significant foreign currency exposure — including IT services exporters, pharmaceutical manufacturers with global sales, and conglomerates with cross-border M&A activity — are deploying AI-powered treasury management systems that can model FX exposure across multiple currencies and maturity profiles in real time, generate optimised hedging recommendations, and monitor market conditions continuously to identify windows for advantageous hedging execution.',

At Infosys, which manages USD revenue exposure of over USD 18 billion annually, the treasury function has integrated machine learning models into its hedging decision process, using historical exchange rate data, macroeconomic indicators, and proprietary signal analysis to inform hedging strategy. The models do not make autonomous hedging decisions — those require human authorisation — but they inform the decision-making process with analytical depth that the treasury team's manual analysis capacity could not replicate.',

The CFO as Architect of AI-Enabled Finance

The CFO's role in AI adoption in the finance function is not merely that of sponsor or approver. The most effective CFOs are active architects of the AI-enabled finance function — making specific choices about which processes to automate, which capabilities to develop internally versus procure externally, how to redesign the finance team's roles and skills to complement AI capabilities, and how to ensure that AI-generated analysis is appropriately validated and governed before it influences consequential decisions.',

This architectural role requires CFOs to develop a working understanding of AI capabilities and limitations that is deeper than the executive briefing level. A CFO who does not understand the difference between a deterministic rule-based system and a probabilistic machine learning model is not equipped to make sound governance decisions about when AI outputs should be trusted, when they should be validated, and when they should be overridden by human judgment. This is not a technical requirement — it is a governance requirement, and it belongs squarely in the CFO's domain.

The workforce implications also require CFO attention. As AI automates the routine analytical and data management tasks that have historically provided the entry-level roles through which finance professionals develop their skills, the pipeline of finance talent faces disruption. How does a junior finance analyst develop the judgment and business understanding that qualifies them for senior roles, if the routine work through which that judgment was traditionally developed is now performed by AI? This is a talent development challenge that India's leading CFOs are beginning to address through deliberate redesign of early-career finance rotations, creating roles that combine AI system oversight with direct business partner engagement from the earliest stages of a finance career.',