I started my career at Wipro, on a team called Mid Market, focused entirely on new logo acquisition across our full gamut of IT services. Sales moved fast. Delivery could not keep up. The reason was simple: our delivery structure was built on specialist manpower, headcount and productivity. And the economics of that structure had a built-in problem. The reward against the delivery effort was never great to begin with, and it diminished further with every new customer we brought in. The more successful we were at selling, the thinner the returns on serving.

We all eventually moved on to different business units, but the promise of that segment, and the thrill of actually solving it, stayed with me.

When we started Scogo, the idea was to build a delivery structure that businesses could subscribe to just-in-time, and be charged for the outcome rather than for the headcount behind it. We managed to change the chargeback mechanism. Customers paid for results, not resources on paper. But the fundamental unit of delivery underneath it all was still resources. We had changed how the bill was written. We had not yet changed what was actually being billed.

That remained true until around two years back, when ChatGPT launched. It dawned on us that we could now build AI agents to complement human resources and rebuild the IT service delivery layer itself, so that it was truly outcome-oriented not just on the customer-facing side, but on the supply side as well. Outcome-based pricing had solved half the equation. This was the other half.

That distinction, I have come to believe, is the real story of the entire managed services industry.

An Industry Built on an Invisible Ceiling

For decades, managed IT services and IT concierge models have operated on an unspoken assumption: that true, exceptional service is only economically viable for a narrow slice of the market. Fortune 5000 companies. Maybe the next 10,000 to 20,000 largest enterprises globally. Everyone else has had to make do with something lesser, or nothing at all.

This was never a deliberate choice by any single provider. It was a structural outcome. Delivering genuinely excellent managed services has always required expensive, skilled people layered on top of expensive, licensed software. That combination only pencils out at scale, and only for customers who can absorb the cost. So the industry drew an invisible line, and most of the world's businesses fell on the wrong side of it.

But here is the part that deserves more scrutiny than it usually gets: even the companies on the right side of that line have been underserved.

Two Kinds of Exclusion

The obvious exclusion is economic. The vast majority of businesses, small and mid-sized companies that make up the backbone of every economy, have simply never had access to the kind of IT concierge experience that larger enterprises take for granted. Not because they would not value it. Every business wants a trusted expert to turn to, a steady hand managing infrastructure, someone who takes real ownership. The demand has never been the question. Affordability has.

The less obvious exclusion is structural, and it affects even those who do pay for managed services today. Look closely at how the traditional model actually makes money, and a quiet misalignment emerges. Managed service providers have long treated third-party software as a margin business. Sell someone else's platform, load a commission on top, and that becomes the economic engine of the relationship.

This sounds harmless until you follow the incentive to its logical end. If a provider earns its margin by reselling software, it has little reason to ask whether that software was the right choice in the first place. Why consider an open-source alternative, or a leaner in-house build, when neither generates the same commission? The provider hands the customer a menu of vendors to choose from and says, in effect, you pick the software, we will run it for you. The responsibility for the single most consequential decision, what technology actually underpins the business, gets quietly outsourced back to the customer, who is usually the least equipped to make that call.

That is the deeper failure. It is not just that managed services excluded most of the market on cost. It is that even where the model reached, it rarely asked the one question that would have made it genuinely valuable: what is actually the best way to solve this problem, regardless of who profits from the answer.

What a Real Concierge Looks Like

We are watching this exact shift unfold in other parts of our lives right now. AI-powered concierges and assistants are becoming part of the mainstream expectation across industries, not because they are novel, but because people have quietly always wanted this: a knowledgeable, responsive, personal layer between them and a complex system. The ideal has never changed. What has changed is that AI is making the ideal affordable at a completely different scale.

This is also, finally, an answer to the problem I saw at the very start of my career. The reason delivery could never keep up with sales, the reason margins thinned with every new customer, was that the unit of delivery was a person, and people do not scale the way subscriptions do. Outcome-based pricing was a step forward, but it was still a financial wrapper around a resourcing problem. AI changes what sits underneath the wrapper. It lets the unit of delivery itself become something that scales, expertise made available on demand, rather than expertise made available one hire at a time.

Applied to IT, this means something specific. A real concierge does not start with a product to sell. It starts with a problem to solve, and then genuinely weighs every path to solving it: building it in-house, using an open-source layer, or subscribing to a professionally built piece of software. Each option evaluated on its own merits, for each specific use case, with no thumb on the scale from a commission structure. That is a fundamentally different posture than picking from a preferred vendor list.

And this is where AI changes the economics in a way that matters. The expensive part of managed services was never the expertise itself. It was the cost of making that expertise available, consistently, at the moment a business actually needed it. AI collapses that cost. It lets a small business access the same quality of judgment, monitoring, and responsiveness that was once reserved for enterprises with dedicated IT departments and premium contracts. The relationship, a human to refer back to, a mechanism you trust, remains intact. What disappears is the price tag that made it exclusive.

Toward a Predictable Idea of IT Spend

There is a version of this future worth imagining concretely. Today, IT spending as a share of revenue varies wildly and unpredictably across companies of similar size, largely because so much of it is shaped by which vendors got a foot in the door rather than by what the business actually needed. If managed services genuinely start from the customer's problem rather than a vendor's product, that should change. IT spend should start to look more like a predictable, comparable line item across peers in an industry, the way other well-understood costs of doing business already do.

That, ultimately, is the test of whether this shift is real. Not whether managed services becomes cheaper, though it will. Not whether more businesses can access it, though they should. But whether the provider sitting between a business and its technology is truly optimizing for that business's outcome, or still just running someone else's margin business with better software underneath.

The question I first ran into as a young sales rep at Wipro is the same question the entire industry is being asked today, just at a much larger scale. AI has finally made it possible to answer it honestly, for everyone, not just the top of the market. The industry now has to decide whether it wants to.