AI in Finance: Automating Skills, Redefining Roles

If you have been following the debate on AI in finance, you will know that it is not really a debate; it is a symphony where every instrument seems to be playing its own tune. Over the last few years, there has been no dearth of papers that have been published on how AI and digital tools are reshaping the CFO function. Some trumpet AI as the ultimate superpower for finance teams, others caution against overreliance, while a few claim that the disruption is overhyped and slow in materializing.

The optimism, contradictions, and occasional theatrics in these writings make for fascinating reading. For every headline that warns of job losses, another pops up predicting finance roles will increase in number. Dashboards and tools are hailed as miracle-makers in one report, only to be dismissed as distracting noise in another. In short, the literature is a mixed bag—but an instructive one.

The future of finance with AI is less about numbers adding up, and more about humans trying to agree on what they mean. 

From automating processes to automating skills

To make sense of the noise, let’s trace the evolution of AI in finance. In the beginning, the spotlight was on automating processes. The rise of ERP platforms and SaaS-based tools promised transaction speed and consistency. But here’s the irony: while processes were streamlined and automated, the size of finance teams did not really change. Over the last 10 years, even for the top quartile companies, the size of finance function per $ 1 B of revenue has moved by just 2 FTEs. Specialized new roles proliferated, almost as if automation had given birth to even more complexity.

We have since crossed into the next phase: the automation of skills. Any skill( human or machine !) has  four fundamental components:

1. Acquire information (absorbing data and inputs)

2. Analyze (decoding and interpreting for decisions)

3. Act (executing steps reliably)

4. Adapt (learning and improving through feedback)

Different waves of technology have chipped away at these parts. RPA scripted the “Act” piece. AI modeling aided “Analyze.” Generative AI and natural language processing have dramatically improved “Acquire,” allowing information to flow seamlessly from both structured and unstructured sources with simple natural language prompts. Machine learning powered the “Adapt” function by enabling autonomous feedback loops.

Now enters Agentic AI—a stage where all four components can be orchestrated together. Earlier, automation picked at the edges, driving incremental savings. Agentic models threaten to replace complete job skills, not just fragments. Finance, long accustomed to living in the world of half-percent efficiency gains, suddenly faces the possibility of entire roles being reframed.

Also Read: Trust in the trust: Governance and Investor Protection

The puzzle of productivity

That said, individual productivity gains do not always translate into enterprise-wide impact. The paradox is unmistakable: employees are busier and more tech-enabled than ever, yet organizations remain plagued by decision delays. EY research shows finance teams still spend 45% of time on mundane tasks and only 20% on high-value activities despite tech adoption.

The problem is less about toolsets and skillsets, and more about mindsets. Finance transformation, once inward-looking (“how do we close faster?”), must now be outward-looking: how does finance create value across the enterprise?  The north star is speed—decision velocity. A company that can see cash movements in real time, forecast margin shifts instantly, and respond to shocks faster than competitors, will always carry the advantage. 

Designing the AI-ready finance function

91% of CFOs express concern over GenAI governance, though 80% have or are building frameworks. Against this backdrop, what does an “AI-centred finance function” look like? A few design tenets stand out:

• Go back to first principles: Don’t just digitize legacy processes; redesign them. A persona-centric and tool-agnostic approach reduces unnecessary process steps while simplifying the finance value chain.

• Efficiency is baseline; quality is focus: Automation ensures the basics. But finance must now differentiate by enhancing capital allocation, improving costing accuracy, and providing sharper commercial insights. In other words: follow the money—literally. Real-time visibility of cash inflows and outflows is finance’s new superpower. 

• Industrialize production, personalize consumption: Finance must behave like a manufacturer of reliable, standardized data while tailoring consumption for varied audiences—be it regulators, auditors, business planners, or boards.

• Model performance rigorously: Driver-linked models that connect operational and financial data are no longer optional. They enable automated causality detection and sharper scenario planning.

• Build guardrails in, not around: With increasing autonomy, AI systems demand continuous control monitoring. This isn’t just compliance—it is about keeping the conscience of the enterprise intact.

When these principles converge, finance ceases to be just the scorekeeper and becomes the story-shaper of the enterprise.

Also Read: Ethics and Financial Discipline Must go Hand in Hand to Foster Trust: Vinay Dhawan, CFO, SNHC

Not AI versus humans, but AI plus humans

One of the biggest myths is that AI in finance is a zero-sum game—machines replacing people. The truth is subtler. AI expands what is possible, often creating entirely new capabilities. Workday and EY studies find 94–99% of CFOs believe AI delivers clear benefits, shifting concern from ‘if’ to ‘how’. The question worth worrying about is not “What can the robots do?” but “What will humans do? This is the unsolved riddle. We are writing long whitepapers about the autonomous future but very little about the future responsibilities of finance professionals. Until this asymmetry is addressed, anxiety within the function will remain.

Finance archetypes of the future

To get unstuck, it helps to think in new archetypes rather than old titles. The finance team of the future may include:

• The Sutradhar (storyteller): weaving narratives from data, helping leaders see the story behind the numbers.

• The Trust Keeper: reinforcing the conscience of machines, ensuring integrity and authenticity of data.

• The Scenario Builder: equipping organizations with the right dictionary of questions for uncertain times. And help drive agility.

• The Gita Bearer: stewarding codes of conduct, ensuring corporate values , laws of the land and ethics anchor financial decision-making.

If today’s CFOs are worried about job descriptions, perhaps tomorrow’s will smile at job epithets. And if the finance function has always been accused of being too serious, these new roles may finally allow it to loosen its tie—just a little. 

Also Read: Specialisation over scale: Workforce planning in GCCs in an AI Economy

Humans at the centre of design

AI is changing the very DNA of finance. From automating acts to automating analyses, and from simplifying processes to reimagining entire skill cycles, the trajectory is accelerating. Yet the most critical work is still unfinished: designing the human role in an AI-shaped world.

AI will not replace finance; but finance teams that fail to reimagine their role risk being replaced by irrelevance. The future does not belong to the robots, nor to the humans alone, but to the audacious few who can orchestrate the two into a lasting partnership.

In the grand balance sheet of life, it’s not humans versus AI, but the synergy column that matters most.