Research overview
Redesigning a labour-intensive business for the AI economy
The starting point
In a labour-intensive services business, revenue is a function of headcount. The unit sold is human time, priced per agent, per hour or per interaction. Technology in that model has historically been a cost lever rather than a revenue line: efficiency gains are real, but they are recovered by clients at renewal. Every previous wave of automation left that structure intact.
AI breaks the symmetry. The volume of routine interactions falls faster than commercial terms can be renegotiated, while the residual mix becomes harder, longer and more expensive to serve. Price and cost move in the same direction at the same time, which is unusual, and it removes the assumption the model rests on — that improving delivery is always a defensible use of capital.
What the work covered
The project treated this as a business-model problem and worked through it in four parts: what could be sold that is not time; how such revenue would be priced and contracted; where human capability remains economically differentiated; and what sequence would move a large installed base from one architecture to the other without breaking the base.
The product work focused on what an intermediary can own that a client cannot assemble alone — proprietary interaction data, ownership of the end-to-end workflow rather than a step inside it, regulated and accountable delivery, and the last mile where a problem is actually resolved rather than routed. The pricing work moved from per-seat and per-interaction toward outcome, resolution and licence-based structures, which changes not only the revenue line but who carries performance risk.
On the human side, the question was not which tasks survive automation but which situations make accountability itself the product: judgement under ambiguity, escalation and trust, regulated or brand-exposed moments, and the supervision, correction and training of the systems themselves. Those are not residual tasks; they are the part of the value chain that keeps commanding a price.
The financial consequence
The transition problem is the hardest part. The new revenue lines start small, grow quickly and carry different margin behaviour; the legacy base is large, still profitable and deflating. Revenue per employee decouples from headcount, gross margin changes shape, and capital moves from capacity to product. For a period, the reported financials describe a business that is disappearing more clearly than the one being built — which is a communication and capital-allocation problem as much as an operational one, and the point at which this work connects to the valuation question I looked at later.
The broader question
Whether a business built to sell capacity can learn to sell outcomes before its capacity is repriced — and what observable evidence would distinguish a transition that is actually happening from one that is being described.
About the author
Amine El Afrit is a Casablanca-based finance executive working across strategic finance, capital allocation, business transformation and operating-model design. His independent research explores AI-driven business-model change, markets and decision systems.