The new economics of enterprise AI with Berger Paint’s Partha Protim Mondal

For nearly two decades, enterprise technology followed a familiar financial logic. Virtualisation reduced infrastructure footprints, cloud shifted capital expenditure to operating expenditures, automation streamlined repetitive tasks. Therefore, each successive technology wave carried the promise that efficiency would eventually outweigh cost.

AI, however, has disrupted that equation. Instead of simplifying enterprise IT economics, AI is introducing an entirely new layer of investment. GPU-intensive infrastructure, soaring cloud consumption, data engineering, governance, cybersecurity and model oversight are expanding technology budgets at a time when boards expect faster and more measurable returns. For CIOs, the challenge is no longer simply deploying AI, but managing its economics.

According to Partha Protim Mondal, CIO of Berger Paints, organisations need to rethink how they evaluate technology investments in an era where productivity, resilience and risk reduction have become just as important as revenue growth.

Cloud’s promise meets operational reality

Cloud computing entered the enterprise as a cost optimisation strategy, yet many organisations are now grappling with steadily rising cloud bills. Now, Mondal attributes this not to flaws in the cloud model but to expanding digital ambitions.

“Cloud computing was initially positioned as a cost-optimisation strategy, primarily through the reduction of capital expenditure. However, as enterprises accelerate digital transformation, many are experiencing rising cloud operating costs. The key driver is rapid consumption growth, as organisations expand digital platforms, analytics capabilities, customer-facing applications and AI-driven initiatives.”

The problem extends beyond higher consumption. Idle development environments, unused storage, disaster recovery infrastructure, overprovisioned resources and data movement across hybrid and multi-cloud environments all contribute to mounting operational costs. AI workloads, with their intensive compute requirements, add another layer of expenditure.

Rather than viewing higher cloud spending as an unavoidable consequence of digital transformation, Mondal argues that enterprises must strengthen financial governance through disciplined FinOps practices that continuously optimise resource utilisation and spending.

AI’s real cost extends beyond GPUs

Generative AI has dramatically changed enterprise infrastructure planning. What was once primarily an IT capacity exercise now demands strategic decisions around computing power, networking, storage and energy requirements. 

Yet Mondal believes many organisations continue to underestimate where AI spending actually goes. “While organisations often focus on AI platform subscriptions and GPU infrastructure, the most significant costs frequently lie beyond technology acquisition.”

Preparing enterprise data for AI remains one of the largest hidden investments. Cleansing, labelling and organising data, building vector databases and maintaining data quality require sustained effort before AI systems begin delivering value. As deployments mature, enterprises must also invest in governance, cybersecurity, compliance and continuous model monitoring to address risks such as hallucinations, bias and model drift.

“Ultimately, AI economics extend far beyond infrastructure. Sustainable value realisation requires disciplined investments in data, governance, security and operational excellence to ensure scalable, secure and business-aligned AI outcomes,” he says. 

Measuring AI beyond revenue

The growing scale of AI investment has intensified scrutiny around returns. Mondal argues that measuring AI solely through revenue generation overlooks much of its enterprise value. “AI should be viewed as a strategic transformation investment rather than a traditional capital expenditure measured solely by direct revenue generation,” he points out.  

At Berger Paints, AI initiatives are evaluated across revenue growth, productivity improvement, customer experience and risk reduction. While revenue-generating AI applications may take time to mature, productivity gains often emerge much earlier through automation, faster decision-making and improved operational efficiency. AI can also strengthen customer engagement while reducing risks associated with cybersecurity incidents, supply chain disruptions, forecasting inaccuracies and compliance failures.

“Boards should adopt an AI Value Scorecard incorporating financial, operational, customer and risk metrics rather than relying on a single ROI measure.” Mondal adds that organisations should expect AI returns to evolve over time, with productivity improvements emerging first and broader competitive advantages materialising over several years.

Convincing the CFO

For many CIOs, securing funding has become as important as selecting the right technology. Mondal believes successful conversations with CFOs begin by reframing AI from an experimental technology initiative into a portfolio of business investments. “The conversation must shift from uncertain AI ROI to the cost of inaction,” Mondal adds.

Rather than pursuing large-scale deployments immediately, he recommends starting with targeted use cases capable of delivering measurable outcomes within months. AI-powered service desks, knowledge assistants and automated reporting can establish confidence before organisations expand into larger transformation programmes.

Equally important is expressing AI benefits in business language. “CIOs must translate AI benefits into business metrics that resonate with CFOs, including productivity gains, working capital improvement, inventory optimisation, revenue uplift, error reduction and risk avoidance.”

The emphasis, he says, should remain on disciplined governance, phased investments and clearly defined success metrics rather than technology enthusiasm.

IT budgets are changing, not shrinking

Predictions that AI will significantly reduce enterprise technology spending may prove premature. “To be very honest, AI is unlikely to reduce total IT spending in the near term; instead, it will fundamentally reshape how technology budgets are allocated.”

In the short term, organisations must simultaneously fund AI infrastructure, cloud platforms, cybersecurity, governance, specialised talent and existing enterprise systems. As a result, technology spending is more likely to increase before efficiencies begin to emerge.

Over time, however, budgets are expected to shift away from routine infrastructure management and manual operations towards intelligent automation, data engineering and AI-driven operations. For manufacturers, these investments could translate into predictive maintenance, better demand forecasting, supply chain optimisation and improved operational resilience rather than immediate reductions in IT expenditure.

Sustainability remains part of the AI equation

The rapid growth of agentic AI has raised concerns about rising energy consumption and the environmental impact of increasingly compute-intensive workloads.

Mondal acknowledges these challenges but argues that sustainability is evolving rather than being sidelined. “The rapid rise of Agentic AI has raised concerns that sustainability may be taking a backseat… However, sustainability will stay; rather, its scope is expanding beyond environmental concerns to include economic, operational and digital sustainability.”

He believes AI itself can become part of the solution by improving cloud utilisation, optimising energy consumption, extending hardware life and enabling more efficient manufacturing operations through predictive maintenance and better resource planning.

The economics of enterprise technology are becoming increasingly complex. Cloud no longer represents a straightforward cost-saving exercise, while AI demands investments that extend well beyond infrastructure into data, governance and operational resilience. For CIOs, the competitive advantage will lie not simply in adopting AI faster than their peers, but in building the financial discipline and governance required to sustain it over the long term.