The Application Management (AMS) contract you've renewed for fifteen years is losing its rationale, and the reason comes down to cost. Artificial intelligence changed the price of building software and made the assumption that propped up traditional AMS obsolete. In its place, another discipline is gaining ground: business process consulting. Renewing classic AMS in 2027 means paying for a service identical to the one you bought in 2020, while the market already operates on a different cost logic.
What business process consulting is
Business process consulting examines each of a company's processes and decides, case by case, the best way to solve it among the available technology options. It looks at the real process, not the legacy system, and applies three criteria at once: technical, economic, and business experience. Its output is a well-founded decision, made with vendor neutrality.
The work starts with mapping the process end to end and ends in a recommendation backed by numbers. In between: charting the current flow, diagnosing where the standard covers things well and where it falls short, modeling the cost of each alternative, and choosing the path. The decision may keep a process in SAP, move it out to a custom application, replace it with an external service, or orchestrate it with AI agents.
The market has already priced in the shift
In January 2026, SAP's stock fell nearly 15% in a single session, its steepest one-day drop since October 2020, after its cloud revenue guidance for the year came in below what the market expected (Reuters). Measured against its high over the prior twelve months, the stock was down close to 40%.
ServiceNow lost roughly 17% in April following its first-quarter results, its worst day on record, and from its all-time high it fell close to 55% by February. Salesforce gave back about a third of its value in six months.
These declines aren't cyclical noise. The market is testing a concrete hypothesis: when an AI agent executes on its own the tasks that once required per-user licenses, the per-seat pricing model comes under pressure. Betting your entire operation on a single vendor today is a position that Wall Street has begun to question by moving capital.
Building software is no longer expensive
The price of building software has collapsed. What used to take a SAP developer a week now gets done in a day or two. Code generation, technical documentation, integration mapping: everything sped up by a factor of ten to a hundred. Pure technical skill, on its own, is no longer an advantage.
The consequence cuts two ways. When the SAP standard covers a process well, building something on top of it loses its justification. When it falls short, there's no longer any reason to put up with it: today it's viable, on both timeline and cost, to stand up a cloud application that extends S/4HANA with the experience the business needs, without touching the core. Two years ago, those timelines and costs weren't realistic.
From two options to six
Three years ago, the question "how do I solve this process?" had two realistic answers: solve it inside SAP, through standard configuration or Z development, or buy specialized software. The rest existed only on paper. Building custom was expensive and slow, modernizing a legacy system was a multi-year project, integrating meant writing your own code, and AI agents weren't around yet.
Today the same question has six viable answers:
- Solve it inside SAP. Standard configuration or Z development. It's still the best answer when the process is generic and the standard covers it, though it's no longer the automatic choice.
- Specialized software. Best-in-class for CRM, human resources, planning, logistics, or pricing, with growing maturity.
- Custom cloud application. Viable today with AI and modern architectures, for the processes where the business truly competes and no standard does it justice.
- Modernize the legacy system. When what was built is holding operations back but the business logic is still sound.
- Smart integration. Integration platforms (iPaaS), APIs, and modern orchestration. The glue between systems is no longer custom code.
- AI agents as a cross-cutting layer. They amplify the other five and automate tasks that used to depend on per-user licenses.
The most common mistake happens before you even choose: it's sticking with the usual option without evaluating the other five.
How you choose: three criteria at once
The question is no longer "how do I configure SAP for this?" Now it's "which of the six options fits this process?" And you answer it with three dimensions of analysis at the same time.
Technical criterion. How well each option solves the problem, what technical debt it leaves behind, how it integrates with the rest of the systems, and how maintainable it will be in five years.
Economic criterion. What it costs to build, run, and evolve each option over three to five years. The real total cost of ownership, with licenses, infrastructure, talent, and debt included. What frees up capital and what keeps it tied down.
Business experience. Which option gives the end user the experience the process demands, which one speeds up the cycle, and which one enables a competitive edge that the others block.
No single dimension decides on its own. A cheaper option that generates technical debt ends up costing more. A technically elegant option that users don't adopt solves nothing. The decision lives at the intersection of all three.
Why traditional AMS no longer cuts it
An AMS set up as incident management resolves symptoms. Every closed ticket, though, is information about how the business operates: what the standard covers well, where it falls short, which critical processes are held together with temporary patches. That information exists in every AMS. In most, it gets thrown away.
The AMS that makes sense today handles each incident twice. The first time, it resolves the user's case within the service level agreement (SLA). The second time, it asks whether the problem will come back and what would keep it from returning. The first response is support. The second is continuous consulting. Without that second layer, tickets reproduce endlessly and the company pays twice: for the license and for the symptom.
What kind of profile does this work
This service calls for a different profile than the classic SAP consultant. Someone who understands the process end to end, knows the six options with real depth, can model the economics of each one, and doesn't answer to any vendor's sales quota. Someone who can say "this stays in SAP," "this moves out to a cloud application," "this gets replaced by an external service," "this gets orchestrated with agents," and defend each decision with all three criteria at once.
That's business process consulting. It works on the process, not the system. It's neutral toward technology. And it uses AI as part of the method, at every stage, from discovery to cost modeling.
What's at stake over the next two years
Over the next two years, companies' technology decisions are going to diverge far more than they did over the past decade. Those who evaluate case by case, with a process mindset and no vendor reflex, will run on systems that are lighter, easier to change, and cheaper. Those who keep buying traditional AMS on the assumption that SAP solves everything will keep paying licenses for capabilities they no longer use and piling up technical debt on processes that should have left the core years ago.
The first step isn't to hire anything. It's to take your five most expensive processes to operate and ask, for each one, which of the six options fits. That single conversation usually reveals where you're overpaying.