.png)
The End of FTE-Based GBS and What Replaces It
- 13 hours ago
- 4 min read
For more than a decade, the logic of Global Business Services has been close to mathematical. Identify a process that costs a certain amount when it sits in the headquarters of a multinational, move it to a location where the same process costs a fraction of that, manage the predictable quality trade-off in the early years, and capture the difference as savings. That equation built the shared services industry, and it has worked so reliably for so long that most of the operating model, the metrics, the management structures, and the career paths inside GBS were all designed around it. We've spent the last six months in conversations with GBS leaders across the region, and most of those conversations are still being held inside that frame, even though the frame itself has started to fall apart.
We've written about what an honest AI readiness assessment actually finds, and what comes next when an assessment lands on the desk and the picture sitting in the deck is uncomfortable. Those articles operate inside the existing model. This one is about something more structural that we think most GBS leaders can already feel even if the conversation in their organisation hasn't named it yet, which is that the operating model the function was built to deliver is becoming obsolete faster than the planning cycles are designed to accommodate.
The question for any GBS leader sitting with their 2027 planning right now is not really how to make the existing operation cheaper or more efficient with AI. It is what the function is actually for, once the labour-arbitrage premise that built it no longer carries the weight it used to.
The cost equation has already inverted, even if the budgets haven't
A junior analyst running structured invoice work, or matching purchase orders to goods receipts, or reconciling intercompany balances, used to be the cheapest reliable way to do that work, particularly in low-cost delivery locations across Southeast Asia and India. That is no longer accurate for a growing share of the work a typical Finance Shared Services function does. A well-supervised AI agent handling the same volume of structured, rules-based transactions is now cheaper than the equivalent headcount by a margin that is not marginal, closer to an order of magnitude in some processes.
What this changes is that the primary question is no longer where to locate the human workforce. It is who is supervising the agents, what governance sits around them, how exceptions are handled, and whether the design of the work itself has been changed to accommodate that the worker is no longer a person.
The metric for GBS performance is measuring the wrong thing
GBS leaders have been measured on FTE per process, cost per transaction, headcount reduction year on year, and ratio of outsourced to in-house. Those metrics describe how efficiently a labour model is being run. They do not describe what the function is producing, what decisions it is enabling for the rest of the business, or what capability it has built that the business can rely on.
The metrics that map to where the function is heading are different in kind, not just in detail. Cycle time on the work, exception rates and how quickly exceptions are resolved, the proportion of work moving through the function without human intervention, the time it takes for a decision the business is waiting on to be produced. These describe a service rather than a labour pool, and they reveal something useful about whether the function is improving or stagnating.
What replaces FTE-based GBS is not "AI-enabled GBS." It is a redesigned service.
A great deal of what is currently being described inside organisations as AI transformation in GBS is the deployment of AI tools inside the existing FTE-shaped processes, with the existing management structures left intact, the existing metrics still being reported, and the existing team designs still in place. That produces visible productivity gains and it looks like progress on a slide. It is not the actual shift, and the organisations moving fastest are not the ones layering AI on top of the old model.
The actual shift, the one the operating model is moving toward whether or not it has been planned for, is that the team is no longer a team of analysts who use tools to do the work. It is a team of people whose job is to supervise and improve the work that AI is doing, design the governance around it, intervene where the agents fail, and continuously raise the floor of what the system can handle on its own. That is a different job description from the analyst role it replaced, a different career path, a different management structure, and a different way of measuring whether the function is creating value.
The FTE-based GBS that built this industry is ending. Not because it failed, but because the shape of the work it was built to do is no longer the shape of the work that needs to be done. What replaces it does not yet have a settled name, but it is taking shape in the organisations that have started designing for AI as the primary worker rather than the supporting tool. That is the conversation the next twelve months will be about, and the leaders who are early to it will have a meaningful advantage over the ones still optimising the model that is on its way out.
AGOS Asia works with GBS organisations across Southeast Asia on the transition from labour-based service models to outcome-based ones, including how to redesign the operating model, the metrics, and the team structure for AI as the primary worker rather than the supporting tool. Start at agosasia.com.
Comments