BPOs are adapting to AI by shifting from selling seats and per-contact volume toward selling outcomes, AI-plus-human delivery, and management of the AI layer itself. The pace of the shift varies a lot between operators.
A retailer hands its support to a BPO and pays per seat. AI shows up. The simple, cheap, repetitive contacts that filled most of those seats are now handled by automation. The seats are still being paid for. The work they do has changed. This is happening in the outsourcing market right now, and what the BPO does about it is starting to separate the operators.
What people in the field are saying
Nick Clark's Service Matters newsletter has been writing about how BPOs are adapting to AI as a live commercial question. The trade press has the same theme: operators that ignore AI are losing margin; operators that embed AI into delivery are repositioning around outcomes.
What is actually changing for a BPO?
Two things at once. First, the share of volume a BPO handles by human seats is shrinking, because AI absorbs the simple contacts at the top of the funnel. Second, the contacts that remain for human agents are harder, which takes longer per contact and changes the cost-to-serve.
Per-seat or per-contact contracts written before AI no longer match the work the BPO is doing. The buyer is overpaying or the BPO is underpricing, depending on which way the contract was tilted.
How are the strong operators adapting?
They are moving away from selling capacity and toward selling resolution. Instead of "we will staff 200 seats," the commitment becomes resolution rate, quality at handoff, and management of the AI layer alongside the human one. They are also taking on the operational work of running AI agents (monitoring, prompt updates, knowledge maintenance) and pricing that as a service.
What are the weaker operators doing?
Holding to the seat-and-volume model and hoping volume holds. It will not. The cost of being late is that buyers will renegotiate the next contract from the outside, or move portions of their support back in-house.
What should a buyer do about it?
Read the contract before the next renewal. Split the contacts handled into easy and complex. Project how much of the easy layer AI will absorb. Use that projection in renegotiation. Ask the BPO explicitly who runs the AI agent layer, who is accountable for its quality, and what happens to pricing if you adopt your own AI on top.
Related: our field note on what AI automation means for your outsourced contact centre, the field note on whether outsourcing still makes sense with AI, and the use case for answering a customer FAQ.