In the bustling corridors of India’s IT sector, “outcome-based” has become the new mantra. Spurred by the AI revolution, industry leaders are declaring a break from legacy billing models. Clients, they argue, are no longer interested in paying for hours worked or personnel deployed; they want to pay for tangible results.

Yet, the shift appears to be more aspirational than actual. While the language on earnings calls and in sales pitches has evolved, the fine print in most contracts reveals a different story. The majority of AI deals continue to follow the very models they’re meant to replace, compelling clients to ask the uncomfortable but inevitable question: “What’s the outcome?”

The Reality Check: Old Wine in a New AI Bottle

Despite the hype, Indian IT services have yet to meaningfully sell AI solutions as true outcome or value-based deals. The numbers, says Gaurav Vasu, founder of the research firm UnearthInsight, tell a stark story.

“About 95-98% of current AI deals are traditional services with AI solutions layered on, focused on productivity improvement targets,” Vasu noted. He added that even the standalone AI proofs-of-concept (PoCs) are typically small, in the $1-5 million range.

A senior executive at a top-tier Indian IT firm, speaking to AIM on condition of anonymity, confirmed this cautious approach. “IT companies are adopting different pricing approaches as no one is clear on how comprehensively AI will impact IT services,” he admitted.

The sticking point is that not every project has a clear, definable outcome. “Take app maintenance projects. Such projects are sometimes priced on a pod or squad basis, with a fixed price for each pod of developers,” the executive explained. 

Other experiments include bill of materials-based pricing, where separate line items cover services, AI agents and tools. 

For Ankit Bose, head of Nasscom AI, this reflects evolution. He described the shift as a “clear progression, not a sudden shift”. He argues that the industry is moving toward a hybrid model that “combines fixed fees with success-linked bonuses tied to measurable business outcomes.”

The Devil in the Definitions: Outcome vs Value

Part of the confusion stems from the way “outcome-based” and “value-based” pricing are often used interchangeably. While related, the two are not the same. Outcome-based pricing ties fees directly to specific, measurable KPIs. For instance, a 30% reduction in customer ticket resolution time. It is tangible and relatively easy to structure. 

In contrast, value-based pricing is a far more ambitious model, linking charges to broader business outcomes, such as revenue growth or market share gains. This approach is much harder to execute, considering quantifying and attributing this kind of impact is notoriously difficult.

Companies are trying to navigate this spectrum. Kalyan Kolachala, MD of SAIGroup India, said his firm leans towards a value-based approach by focusing on the “productive benefits” AI brings. “If you’re saving X, can we get a small percentage of that X? That way, pricing is tied to outcomes,” he explained.

Hasit Trivedi, chief digital and AI officer at Firstsource, said his organisation is already there. “For us, AI pricing always starts with the ‘what’, what business value can we deliver,” he said, adding that their commercial models are “driven by value”. However, he concedes that “the real challenge is measuring the full impact”.

Why Deals Go Into ‘Cold Storage’

This measurement challenge is one reason many promising conversations are stalling. Vinay Kumar HS, global business head at AidenAI, noted on LinkedIn that deals often slip into “cold storage” due to a lack of clarity. “One of the key reasons for client hesitance is the lack of clarity around the KPIs used to define success, as well as the internal dependencies required to achieve those outcomes,” he explained.

Multiple critical roadblocks stand in the way. A big one is the fear of cannibalisation. As NimbleEdge co-founder Neeraj Poddar pointed out, vendors charging per-seat for human agents face a dilemma where pricing an AI agent per-ticket could undercut their existing, lucrative business. 

Attribution poses another hurdle. Attributing revenue uplift directly to an AI solution is “never straightforward,” Poddar observed, making pure value-based models a tough sell. 

Cost unpredictability and risks adds to the caution. Kolachala revealed that “only when scaled in volumes do clients realise that…costs can be higher than expected,” citing expensive calls to large language models. 

Moreover, the risk of AI “hallucinations” remains a persistent concern, requiring costly human oversight.

A Radical Alternative: Ditch Complexity for a Flat Rate

Frustrated with the messy variables of both legacy and emerging models, some experts are proposing a third way. Saurabh Gupta, president (research and advisory) at HFS Research, argues that both effort-based billing and outcome-linked pricing are becoming obsolete.

“In an AI-led world, effort-based pricing doesn’t just feel outdated, it’s meaningless,” Gupta stated. “Measuring its value in FTE equivalents is like measuring streaming bandwidth in DVDs shipped.”

He pointed to the unpredictable nature of AI costs, ranging from compute power (NVIDIA GPUs), token charges (GPT-4) and storage to fine-tuning and API fees. His solution: flat-rate pricing. 

The model is simple—a single, all-inclusive annual fee for an AI platform within a defined scope. For example, he shared, “$150,000 a year for an AI-led customer query automation platform handling up to 10 million queries.” This approach offers clients cost certainty and incentivises providers to optimise their operations.

Gupta frames this as a “trust-building move”, on his blog. Yet it comes with risks. What happens if a client’s usage explodes beyond the scope, making the contract unprofitable for the vendor? Conversely, what if the client pays a large fee but their usage is minimal? He did not respond to requests for comment on how such pitfalls might be addressed. 

For now, the Indian IT industry remains in a state of pricing uncertainty. 

Everyone agrees the old world of time and materials (T&M) and full-time equivalent (FTE) billing is fading, but a clear, scalable and sustainable model for the AI era has yet to emerge. 

The transition from selling effort to selling outcomes is proving to be the industry’s most complex project yet.

The post Indian IT Caught Between Hours and Outcomes appeared first on Analytics India Magazine.