There is a particular kind of problem that only the ambitious encounter — the problem of having moved so fast that the scaffolding hasn’t kept up. India’s mid-market enterprises know this problem intimately. They were first into the AI arena, first to deploy at scale, first to embed artificial intelligence across core business functions. And now, with characteristic candour, they are among the first to confront what that speed has cost them.
The Freshworks Global Cost of Complexity Report 2026 puts a precise number on it: 27% of the average mid-market AI budget in India is consumed not by innovation, but by what the report dryly calls “complexity overhead.” That translates to an estimated ₹33,000 crore in wasted AI spending annually — a figure higher than the global average of 25%, and a direct consequence of scaling fast across fragmented, often mismatched technology environments. The irony is sharp: India leads the world in mid-market AI integration, with 36% of organisations embedding AI across multiple core business functions — more than double the global average of 15%. Yet the more deeply AI is embedded, the heavier the operational burden it appears to create.
The Audacity Dividend — and Its Shadow
To understand how India arrived here, it helps to appreciate the sheer scale of the ambition. Saket Dandotia, Co-Founder and CEO, Onetab.ai, captures it with a line that doubles as a manifesto: “The world invented AI. India made it useful.”
It is a bold claim. It is also, by the numbers, a defensible one. India leads the world in generative AI adoption at 73%, with coding queries running at three times the global average and data analysis at four times. Enterprise deployment has outpaced the US, the UK, and every other mature digital economy. And unlike markets that have approached AI cautiously — through extended pilots, governance frameworks, and phased rollouts — India’s enterprises largely chose to deploy first and refine later.
But that very instinct — to move, to build, to commit — is precisely what the complexity report is measuring. The cost of going fast is, in part, the cost of going messy. And Rajiv Dingra, Founder and CEO, ReBid, is clear-eyed about what that means on the ground: “Indian enterprises are moving fast, but many are also discovering that AI without integration, clean data, workflows and governance can become another layer of complexity rather than a productivity multiplier.”
The Freshworks data bears this out in uncomfortable detail. Some 88% of Indian mid-market IT leaders report that managing AI complexity has increased their team’s workload — meaning that in nearly nine out of ten organisations, AI is creating effort, not just replacing it. The average Indian mid-market firm now runs 4.6 AI tools; 16% are juggling seven or more. Managing that sprawl is consuming 27% of AI-related working time — time spent not on value creation, but on coordination, validation, and troubleshooting the very systems that were supposed to simplify things.
When AI Becomes the Problem It Was Meant to Solve
Perhaps the most telling symptom of this complexity trap is what happens after AI produces an output. According to the Freshworks report, 85% of Indian IT leaders say AI-generated content introduces noise, errors, or rework into their workflows. The productivity loop, in other words, is broken at the last mile.
Kartik Mehta, CBO and Head of Asia, Channel Factory, sees this not just as an operational headache but as a market-level phenomenon that is reshaping the brand environment. “As India shows signs of fast adoption of AI, we are also noticing an increase in low-quality content,” he observes. And what distinguishes today’s version of this problem from its historical predecessors is velocity. Low-quality content has always existed. What’s changed, as Mehta puts it, is “the velocity and scale of its creation.”
This is the phenomenon the industry has begun calling “AI slop” — content manufactured for monetisation, stripped of intent or craft. For brands navigating a media environment increasingly saturated with it, Mehta argues the challenge is not merely defensive. He introduces the concept of the Unsloppable Brand — one that doesn’t just sidestep low-quality adjacency but actively converts brand safety into brand advantage. The prescription: audit where your signal is coming from, interrogate what you are actually optimising for, and resist the temptation to conflate platform performance with genuine brand presence.
Siddharth Kelkar, Managing Director of India/MENA and Performance Business, AnyMind Group, takes a structurally similar position but frames the problem at the ecosystem level. For Kelkar, the rework and noise are symptoms of a deeper misalignment — organisations that have accumulated AI capabilities without building the connective tissue that makes them function together. “Many businesses have reached a point where they’re spending as much time managing AI as they once spent managing manual processes,” he notes, before adding the crucial qualifier: “and that’s not because the technology isn’t working.”
It is a distinction worth sitting with. The problem is not AI. The problem is the environment AI has been dropped into — and the assumption that adoption alone constitutes strategy.
The ROI Gap Nobody Wants to Talk About
There is a further tension embedded in the Freshworks data that the industry is only beginning to address openly. Some 74% of Indian executives expect AI investments to deliver measurable returns within eight months. The actual deployment timeline, in most mid-market organisations, runs between six and twelve months — meaning the return window and the deployment window are, in many cases, the same window. Programmes are being evaluated, and in some instances discontinued, before they have had sufficient time to generate any outcome worth measuring.
Dingra names this directly. “India’s AI adoption story is no longer about intent — it is about execution.” And execution, in his framing, has a specific meaning: not the act of deploying a tool, but the act of making it work durably within existing business systems, with governance, clean data, and measurable accountability baked in from the start.
The structural barriers to getting there are well-documented in the Freshworks report: 34% of Indian respondents cite system integration complexity as a key obstacle; 31% point to excessive configuration requirements; 30% flag talent shortages. Together, these factors slow the journey from pilot to production — and it is in that gap, between proof of concept and scaled deployment, that most of the wasted ₹33,000 crore disappears.
Maturity Looks Like Simplification
What emerges from both the data and these industry perspectives is a coherent picture of what India’s next phase of AI adoption needs to look like — and it is less about capability accumulation than it is about deliberate simplification.
Kelkar puts it plainly: “The next stage of maturity will be less about adding new capabilities and more about simplifying the environments those capabilities have to operate in.” Success, in his view, will come from building connected ecosystems — AI deployed in ways that fit into existing workflows, teams, and decision-making processes rather than layering on top of them.
The data reflects a market already shifting in this direction. In India, 93% of IT leaders now say they prefer AI solutions with built-in workflows over those requiring extensive configuration. Some 44% rank workflow integration as their single top priority. The appetite for complexity, quietly and without fanfare, appears to be fading.
Dandotia, characteristically, frames the trajectory in larger terms. The 73% adoption figure, he argues, is not the ceiling — it is the floor. “Per capita, most of India hasn’t meaningfully started yet. The 73% is the early wave. The engineers, the entrepreneurs, the first movers. Behind them sits the largest untapped AI-ready population in the world — young, digitally connected, and hungry.” For him, what India is building right now is not a trend reaching its peak. It is a foundation being laid for a transformation of a scale the global technology industry has not witnessed in a single market, in a single generation.
The Reckoning — and the Opportunity Inside It
These are not comfortable conversations. They require Indian enterprises to acknowledge that moving fast created real costs — costs now borne by IT teams stretched thin, brands navigating content environments they didn’t design, and finance functions staring at a ₹33,000 crore gap between spend and outcome.
But discomfort, here, is a marker of maturity. The markets that are not having these conversations are the ones that haven’t moved far enough to need to. India’s complexity problem is, at its core, an adoption problem — one that only countries leading the field are positioned to experience.
Dingra captures the stakes with clarity: “The winners will be companies that don’t just adopt AI, but operationalize it across teams, data, processes and customer journeys.” Mehta adds the brand dimension: those who don’t just avoid the noise but build brands strong enough to cut through it. Kelkar maps the architectural requirement: ecosystems, not tool stacks. And Dandotia holds the long view: a country not at its ceiling, but at the beginning of something far larger.
India’s mid-market has already answered the question of whether it can lead the world in AI adoption. The question it is now answering — in real time, at significant cost, and with characteristic energy — is whether it can lead the world in making AI work. The evidence suggests it will. But the path there runs not through more tools, more deployments, or more ambition. It runs through the harder, quieter work of making what already exists actually function.
