Enterprise AI adoption hits monetization wall as ROI questions mount

Enterprise AI Spending Outpaces Revenue Generation

Companies are discovering that heavy AI spending doesn't automatically deliver financial returns. Satish Viswanathan1

1

, former global data and AI lead for travel at Accenture and author of The Weight of Intelligence, warns that organizations treating token consumption as a measure of AI success are missing the bigger picture. "Just because you consume tokens doesn't automatically mean there is a bank and you can start having a teller to withdraw money," he explains. Some companies even specified token consumption commitments in hiring packages during 2024, treating it as a capability metric rather than focusing on measurable business outcomes from AI spending1

1

.

Rafi Katanasho2

2

, Asia Pacific CTO at Dynatrace, echoes this concern: "There's a lot of tokens being consumed here. But what is the outcome? What are the business outcomes from that investment going on?" Boards and chief financial officers are increasingly demanding clarity on AI ROI as companies redirect budgets from traditional technology investments toward AI initiatives2

2

.

Challenges Enterprises Face in Monetizing AI

Viswanathan introduces the concept of "enterprise metabolisation" to explain why access to capable AI models doesn't guarantee productivity or profitability. Organizations face multiple constraints including technology debt, people debt, process debt, and organizational-structure debt that limit their ability to absorb and use AI effectively1

1

. Model capability alone means little without proper data architecture, system integration, human processes, and governance frameworks.

Even if artificial general intelligence became available tomorrow, Viswanathan argues, not every enterprise would benefit equally. "Still, every enterprise is unique," he notes, emphasizing that technology debt and organizational readiness determine how much intelligence a company can actually deploy1

1

.

Scaling Autonomous AI Agents Remains a Bottleneck

Indian enterprises face particular challenges with scaling autonomous AI agents. While Deloitte India's March 2026 State of AI in the Enterprise report found 40% of Indian respondents reported significant or full AI usage compared with 28% globally, deploying AI differs fundamentally from granting agents authority to act without human intervention2

2

.

Katanasho identifies the core problem: "India has largely solved for access to AI. However, the bottleneck is the confidence in delivering AI in a production environment, at scale." Autonomous agents can fail silently, delivering poor business outcomes without triggering system alerts. In India's digital payments ecosystem, even 99.9% accuracy could create meaningful challenges when applied across massive transaction volumes2

2

.

Many companies are adopting a "bolt-on AI" approach, automating smaller process segments instead of redesigning entire workflows around autonomous systems. Indian banks experiment with AI agents for know-your-customer processes and credit validation, but remain cautious about extending automation to complete business processes where error consequences prove significant2

2

.

Productivity Gains Don't Equal Financial Returns

Viswanathan challenges the assumption that productivity improvements automatically become savings. When an employee saves an hour using AI, that time doesn't necessarily translate into financial gain. "We are not in this factory model where the moment we go at 9 am when you punch your card, you're invariably working," he observes. Knowledge workers may use freed time for personal activities rather than additional productive work1

1

.

Technology services companies face similar challenges with enterprise AI adoption. If AI enables faster project completion, clients may demand lower contract prices, preventing providers from retaining efficiency gains. The ROI of AI investments depends on how organizations redirect saved time—whether through new responsibilities, skill development, or expanded project scope1

1

.

Accountability Issues Threaten AI Monetization

Katanasho emphasizes that "autonomy requires authority. But authority without the truth is very dangerous." Companies need independent systems to observe agent interactions end-to-end and hold outcomes accountable, not just individual agents. Monitoring agents in isolation won't reveal where mistakes occurred or how they affected final results2

2

.

In banking, fintech, and healthcare, silent failures in automated processes can affect sensitive data and service availability. Organizations must track how agents interact with each other, which models they use, and which enterprise systems they access. Without this visibility, companies struggle to identify unexpected agent behavior before it impacts business outcomes2

2

.

The Path Forward: AI-Native Process Redesign

Both experts see greater value in AI-native projects that fundamentally rethink business processes rather than adding AI to existing systems. Katanasho states: "The ultimate value, in my view, from agentics specifically, is transforming that business process altogether. Process reengineering through AI." He compares this transition to cloud computing adoption, where businesses that redesigned operations around the technology captured greater benefits than those merely experimenting2

2

.

Viswanathan notes a strategic pause in enterprise AI adoption. Companies aren't abandoning the technology but are pursuing fewer use cases simultaneously and reconsidering investment scale. "We've already seen very, very smart pausing," he describes, linking this reassessment to pressure on chief executives to demonstrate returns. Organizations still need to determine "what is the use case, what is the area of intervention and how to make money out of it"1

1

.