Twelve years ago, the company outsourced its billing. The decision was right at the time: the activity was standardized, rule-based, and significantly cheaper with a specialized service provider than in-house. Now the contract is up for renewal, and a pilot project is challenging the old certainty. An AI-supported process performs the same work in-house, faster than the service provider and below their quoted price. Added to this is something that doesn’t even appear in the quote: client data no longer leaves the company, and knowledge of the internal billing process returns to where it belongs. Outsourcing was the right answer for twelve years. It’s just that the question has changed in the meantime. <\/p>\n\n
\nThe outsourcing of standardized work was based on a labor cost advantage. AI is causing exactly this advantage to shrink. As a result, the boundaries of the company are shifting again, this time inward. <\/strong><\/p>\n<\/blockquote>\n\n
An energy provider I advised<\/a> on this decision recalculated the figures before the contract renewal. The result: AI-supported in-house production was below the provider’s offer, with shorter lead times and full data sovereignty. In the end, the company even initially decided against reshoring, but the mere existence of a verifiable alternative fundamentally changed the negotiation. For the first time in years, the company faced the provider as a client who had a real choice, and the renewal price reflected that. <\/p>\n\n
What is happening here is not an isolated case, but a structural shift. Three levers turn this into an executive decision. <\/p>\n\n
The calculation has reversed<\/h2>\n\n
The wave of outsourcing in recent decades followed a clear logic: standardized, rule-based work was cheapest where wages were lowest or where a provider bundled the volumes of many clients. AI is now targeting exactly this type of work. Erik Brynjolfsson, one of the leading economists of digitalization, described the core of this development early on: machines do not take over professions, but tasks, starting with uniform, rule-based ones. Once these tasks are digitalized, their marginal costs drop toward zero: every further invoice, every additional document check costs almost nothing. The labor cost advantage that justified outsourcing thus loses its foundation, because software has no location advantage. <\/p>\n\n
At the same time, the costs that have always spoken against outsourcing and were rarely quantified remain: managing the provider, coordination across the interface, dependence during price negotiations, and the creeping loss of internal process knowledge. When the cost advantage shrinks and these burdens remain, the decision tips. The boundary of the company, which has shifted outward over the years, is moving back inward for rule-based work. <\/p>\n\n