The companies built on tens and hundreds of thousands of billable agents now face a brutal question: can they survive a market that wants outcomes, automation, and flexibility instead of headcount?
CC Photo by Andy Dutton
A Shift Bigger Than BPO
There is a structural shift taking place in the services provided by business process outsourcing (BPO) companies. Much of this work can and will be automated using AI, creating a blended service where humans work alongside automated systems.
This isn’t unique to BPO. Most professional service companies are facing the same challenge right now. Earlier this year, the Harvard Business Review podcast explored how McKinsey has changed their entire business model – moving from charging clients by the day or hour for their consultants, toward partnering with companies and earning from the savings – or new revenue – that their expertise creates.
This extreme change in management consulting has been triggered by a simple fact: tools like Claude can now take global inputs and distill them into a strategy framework in minutes. Why pay a consultant to write something that you can generate yourself?
Accountants are facing the same challenge. Lawyers too. All these professions are finding that they need to move beyond charging for time or headcount, because AI automating the intellectual processes and the connection between value created and headcount is less pronounced.
This is great news for most consumers. If you are paying for tools like TurboTax – plus additional accounting advice – then it’s likely that your next tax filing could be a lot easier. Just ask an automated assistant for advice. But naturally, it’s less good news if you just started out in the accounting career you expected to last for decades.
Applying This To BPO
Now apply this mindset shift to BPO. Most of these companies have always charged for their services based on the number of people providing service from a contact center or work at home. If you have 500 people helping customers then the BPO charges 500 x a monthly fee – which includes the staff cost, technology and infrastructure, and margin.
It’s simple and easy to understand. But if AI starts handling 40% of those customer interactions, what happens?
AI is not free, so a client can’t reasonably demand a 40% reduction in rates. But at the same time, if the BPO is increasingly automating more and more customer interactions, then how can they continue charging just for headcount?
The Market Already Knows
Every major publicly-traded BPO knows this is a problem. Their investors also know it. Some of the largest players have seen their share price tumble 90% in the last couple of years – the exact period in which generative AI has reshaped professional services.
Imagine if a chunk of your 401(k) was invested in BPOs in 2022. Now you are looking at an investment that is worth ten cents for every dollar you invested. The market has already priced in an uncertain future for the industry.
The challenge is sharpest for the mid-size and largest players. The smaller companies are more agile and can introduce charging models that are more closely aligned with the way that services are actually delivered now.
I won’t single out individual companies here – several of the largest employ tens and hundreds of thousands people, and all of them face this same structural problem. Most of this headcount is billed on time and-material contracts – even as all clients increasingly ask for outcome-based agreements tied to what is actually delivered.
An organization with thousands of people billed largely on headcount has a safe and predictable revenue stream.It knows what it’s getting each month, regardless of what’s delivered. The difficulty is the transition from that model to to one paid based on customer satisfaction or customer engagement – regardless of headcount.
Every Contract… Contested
I expect this will play out contract by contract. As each service contract nears renewal, the BPO will be invited to submit a new operating model – one that reflects how they will use AI to automate some processes and how the remaining human work will be priced. What does the blend look like, and who pays for what?
The result: every BPO contract becomes contestable. There is no more automatic contract renewal simply because a client is happy with the relationship. Every contract will be challenged. Every client will invite the more agile and innovative BPOs to pitch – including ones they’d never previously considered.
Consider it from the client’s side. If your BPO contact renews three years from now, would you wait that long? Do you know how your business will be operating next year – what will have changed? Most leaders are already struggling to keep pace with constant change – OpenAI updates ChatGPT every one or two months. Can your operating model really wait three years?
The Narrative Problem
Part of what makes this one of the biggest problems facing the larger BPOs: analysts, business commentators, and investors are all telling the story of their decline. This is natural – a short seller benefits from talking loudly about BPOs downfall. But the market narrative and the operating reality now reinforce each other.
All BPOs know they must transform. The real question is whether an organization with agents being billed monthly can genuinely become a dynamic outcome-based business – and do it while servicing the debt that built its scale in the first place?
Many of these players carry significant bond and note obligations that require reliable revenue to service. Maintaining that revenue while simultaneously renegotiating every client contract around outcomes is not a small ask. Square that circle.
Scale Was The Moat. Now It’s The Trap
The real question is not whether BPO will survive AI. The concept of BPO will survive – companies will always need partners that can manage complexity, customers, technology, regulation, language, culture, and service at scale.
The existential question is whether the largest BPOs can dismantle the very business model that made them large in the first place. For decades, headcount was the engine of growth and the basis of the pricing model. More agents meant more revenue. More seats meant more predictable contracts. More people meant more power.
But in an AI-enabled, outcome-based world, size can become a trap. A company built to sell hours may struggle to sell results. A business designed around managing tens and hundreds of thousands of people may find it painfully difficult to reward automation, efficiency, and fewer interactions.
The giants may not fail because they ignored AI. They may fail because they understood the future, but could not move fast enough to escape the weight of their own past.