The Suite Changed. Most Partner Business Models Did Not.

A partner logo on the wall is not a business model.

SAP’s public direction is settled: one AI-driven cloud suite, sold as one motion — applications, data and AI together, with Joule repositioned as the operating layer and agents across five domains. Whatever you think of the roadmap language, the commercial logic behind it is not forward-looking. It is already how SAP sells.

Most partner business models were built for a different SAP. Perpetual licences created implementation projects. Complex customisation created multi-year programmes. On-premise landscapes created hosting margins and upgrade cycles. An entire generation of service firms grew profitable on the distance between what SAP shipped and what customers needed.

That distance is closing — deliberately, from the vendor’s side. This piece is about what that means for the people who run partner businesses, not for the people who buy from them.

Why this is a business-model question, not a portfolio question

Cloud standardisation compresses the effort partners used to bill. Embedded AI compresses it again. What used to be a customisation programme becomes configuration plus disciplined extension. What used to be an upgrade project becomes a release the vendor delivers. What used to be a support contract becomes part of a subscription the customer already pays.

The revenue does not disappear. It moves — towards process standardisation decisions, data quality, adoption and change, industry-specific intellectual property, and the continuous work of proving that promised value actually arrived.

Adding “AI services” to the portfolio does not answer that shift. A portfolio describes what you offer. A business model describes what you are paid for, by whom, and why they could not get it cheaper elsewhere. The uncomfortable question for many partner leaders is that their answer to the second question still assumes the old SAP.

Three decisions, one system

The renewal is not one decision. It is three, and they only work together.

Business model. What does your firm sell when billable effort stops being scarce? Outcome-linked services, packaged industry solutions, managed processes, advisory that changes decisions rather than documents them — each is a legitimate answer. “More of the same, delivered harder” is not.

Organisation. Who delivers the new answer? A transformation practice bolted onto a project factory fails predictably: the factory wins every resourcing conflict, because the factory pays this quarter’s salaries. New business needs dedicated people with their own targets — and leaders willing to protect them from the core business’s gravity.

Go-to-market. How do you create demand when the vendor’s own motion changes around you? Referral flows shift, co-sell rules shift, and the midmarket runs through partners by design. A firm whose pipeline is mostly passed-through vendor leads does not own a go-to-market. It rents one.

Decide one of the three without the others and you get the familiar failure modes: a strategy paper nobody staffs, a new unit nobody sells for, a sales message the delivery organisation cannot honour.

Five questions that expose the gap

I have yet to meet a partner leadership team that disputes the shift. I meet many whose operating reality has not registered it. Five questions usually make the gap visible:

Strategy. Can you state on one page what your firm sells in 2028, and to whom — or is your strategy the sum of your current projects?
Ownership. Who owns the new business — named people with dedicated capacity and targets, or “everyone, alongside the day job”? Everyone means no one.
Demand. If vendor-sourced leads stopped tomorrow, what would your pipeline look like in six months? Own demand generation is the difference between a business model and a subcontract.
Sales. Can every seller explain the new suite in the customer’s language — or do your deals depend on the two people who can?
Delivery. Are your delivery skills concentrated where the roadmap is going, or where your reference projects were?

None of these questions is about AI. All of them decide whether an AI-era business model is executable.

Reality check

Two honest caveats.

First, not every partner must become a product company. Deep specialisation in one industry or one process domain is a durable model — arguably more durable than a shallow full-service claim. But it has to be chosen deliberately, with the economics done, not defaulted into because change was postponed.

Second, the vendor relationship still matters. Programmes, certifications and co-sell motions are real commercial infrastructure, and the delivery models they support shape what customers buy. But programmes reward what you already are. They do not decide what you become. That decision does not arrive from Walldorf.

And a transition like this is measured in years, not announcements — in cohorts of people retrained, offerings retired, and margins rebuilt. Any partner strategy that fits in a press release is not one.

My take

The partners that will matter in 2030 are recognisable today — not by what they announce, but by what they have stopped doing. Business-model renewal begins with subtraction: the offering you retire, the revenue you deliberately walk away from, the project type you no longer staff. Firms that cannot name what they have stopped have not started.

The suite has a roadmap. The question is whether your business model has one.