Managed services is a business model in which a provider runs a whole function for a client, such as customer support, IT or payroll, under a long contract with agreed service levels. The provider earns its margin by doing the work more cheaply and reliably than the client could, usually through scale and trained staff.
RewiredAI rewires it: clients can automate the simplest work themselves, while a new line of work, preparing and checking the data that trains AI, opens up for the providers.
How it makes money
A managed services contract is a promise about results, with people underneath. A client hands over its customer support or its back office, and the provider guarantees how fast the phones are answered and how few errors slip through. The margin lives in the gap between what the client would pay to do the work and what the provider pays to do it.
That gap is what AI changes. If software can answer the simple questions, the client may not need to hand them to anyone. But the software needs data, labeled and checked by people, and someone has to do that work too.
The simple work goes home
Concentrix, which runs customer support for large companies, says it plainly in its 10-K: "some lower complexity services currently performed by our game-changers have been and may continue to be replaced by tools deployed by clients." Game-changers is its word for its staff. Its revenue fell 1.2% in the quarter ended August 31, 2026, though it was up 2.0% for the year to date.
The new work
TaskUs shows the other side. Its AI Services segment covers "large language model support and high-quality data labeling services," the work of training and tuning AI models. Revenue from it was $214.2 million in 2025, against $135.1 million in 2024, up 58.6%. Total service revenue rose 19.0%, to $1,183.5 million.
How it fails
A managed services provider fails when its clients stop needing the volume it was built for. Contracts are priced on staff and service levels. When software takes the easy cases, what is left is the harder work, at a different cost, and the provider has to reprice it or shrink.
Our read
The outsourcing deal was always a bet that someone else could do the work more cheaply. AI takes the cheapest work out of the deal altogether: not to another provider, but back to the client, done by software. What remains for the provider is the work software cannot do alone: the hard cases, the exceptions, and the people who train and check the models.
For a provider, that makes the mix the strategy. Revenue from simple volume is the most exposed; revenue from judgment and from AI's own supply chain is where the growth is. For a client, the question to put to every outsourcing contract at renewal is which of its tasks a tool you already own could handle.
Questions
- What are managed services?
- A provider runs a whole function for a client, such as customer support or IT, under a long contract with agreed service levels, and earns its margin by doing the work more cheaply or reliably than the client could.
- Is AI replacing outsourced customer service?
- In part. Concentrix says some lower complexity services its staff performed have been replaced by tools its clients deployed. Its revenue fell 1.2% in the quarter ended August 31, 2026, though it was up 2.0% for the year to date.
- What is AI data labeling work?
- Work done by people to label, annotate and check the data used to train and tune AI models. TaskUs reports it as AI Services, which grew 58.6% to $214.2 million in 2025.
Sources
- Concentrix Corporation (SEC EDGAR), Form 10-K for the fiscal year ended November 30, 2025 (2026-01-28)
- Concentrix Corporation (SEC EDGAR), Form 10-Q for the quarter ended August 31, 2026 (2026-10-07)
- TaskUs, Inc. (SEC EDGAR), Form 10-K for 2025 (2026-03-05)
Every figure and quotation on this page traces to these filings. “Our read” is Stratrix’s judgment, drawn from them. Analysis and data, not advice.
More from Stratrix
“You gave the work away and want it back; what does that cost now?”
See the pack — $499 →The Business Model Wizard
Make the seven choices that define your model, test your numbers against real companies, and meet the rival AI would build against you.
Free · about twelve minutesBuild your Model Thesis →Test a decision, free
Set a live decision against the companies that made it before you.
AcquisitionsThe Distance TestDivestituresThe Divestiture TestDiversificationThe Diversification TestMake or buyThe Make-or-Buy TestWeekly email
Stratonomics, every Tuesday
One data story every Tuesday: what happened, what it means, and the figures you can quote. Free.