USF Convention 2026: the key SAP trends for HR, Finance and IT

This year's central theme came down to one question: what do we absolutely want to keep under our control? For the USF, being sovereign does not mean doing everything in-house, nor rejecting the Cloud. It means keeping the freedom to choose:

• your architecture and where your data is hosted;
• your partners and your contracts;
• your pace of transformation;
• your key skills, in-house.

Innovation and control are becoming compatible.

Good to know

According to the USF 2026 survey, 31% of SAP customers are legally required to use a sovereign cloud and 30% are concerned about it for geopolitical reasons. The topic now reaches well beyond the public sector.

The 7th USF survey, carried out with the Verian research institute among 276 respondents, shows real momentum, but an increasingly tight timeline:

• A two-speed migration: 31% of customers have completed their move to SAP S/4HANA, but 40% have not started yet.
• A deadline drawing closer: ECC maintenance ends in 2030. The more projects bunch together, the scarcer the skills available on the market.
• The Cloud is gaining ground: 36% of customers have already signed an ERP Cloud offer, mostly in private cloud. The main obstacles remain costs, visibility on future pricing and how responsibilities are shared if an incident occurs.
• A more pragmatic approach: brownfield is on the rise (44%, up from 34% in the previous survey) and greenfield now accounts for just 20%. Secure first, transform next.

Key takeaway

Starting the scoping phase now means you can choose your migration strategy, secure the right skills and stay in control of both your timeline and your budget.

SAP sums it up as “humans set the framework, AI executes”. Agents handle tasks, while teams focus on decision-making. Among the key announcements:

• Joule becomes SAP's working interface: a request in natural language, an analysis, a recommendation, then execution.
• 400 agents are announced by the end of the year, and Joule Studio lets you build your own agents without code.
• A reduction of at least 35% is targeted in the duration of transformation projects, thanks to assistants (data migration, testing).

Customer experiences point to the same method: useful AI starts from everyday pain points, not from a catalogue of agents. Several SAP customers shared good practices:

• Identify low-value tasks that slow teams down.
• Put a figure on every use case, including the cost of AI consumption.
• Be ready to drop some: in one project presented, half of the 40 use cases identified were abandoned because they did not create enough value.
• Govern your agents: SAP mentioned a customer where 2,000 uncontrolled “shadow AI” agents had been detected.

Generative AI is included, agents are charged on a usage basis

This is the most tangible announcement for HR departments: SAP now draws a clear line between two levels of AI in SAP SuccessFactors.

1. Embedded generative AI: included in your licences

• Cost: included in SAP SuccessFactors licences from 1 November.
• Scope: all modules (recruiting, Core HR, payroll, onboarding, etc.).
• Examples: automatically drafted goals, improved tone in performance reviews.
• To plan for: activating the features and supporting users.

2. Agentic AI: agents charged on a usage basis

• Cost: agents consume specific SAP credits (AI Units).
• Scope: dedicated agents for certain HR and payroll processes.
• Example: an agent that explains their payslip to an employee.
• To plan for: the credit budget and the choice of priority use cases.

Worth noting: AI innovations reach SAP SuccessFactors first, so SAP SuccessFactors customers are the first to benefit.

SAP payroll has a clear path to the Cloud

Companies running SAP HCM payroll have two options:

• Employee Central Payroll, the target recommended by SAP;
• SAP HCM payroll in private cloud, supported until 2040, if urgency or time management requires it.

A well-prepared migration remains manageable: one project presented took 6 months, 4 consultants and 3 months of parallel payroll runs. The keys: treat time management as a separate workstream, remove custom developments, and plan ahead.

A tool gets adopted when everyone can see what's in it for them

As Christian Clot, director of the Human Adaptation Institute, pointed out: nobody changes their behaviour simply because they have been given information. To drive adoption of a new HR tool, it helps to:

• Show everyone what they gain, not just what changes.
• Accept that teams move at different speeds.
• Build the time actually required from each person into the project plan.

In a multi-site group, the first building block is the least visible, yet everything else depends on it. Making Employee Central, the Core HR of SAP SuccessFactors, the single source of employee data means onboarding, talent or compensation can then be rolled out on a reliable foundation. Payroll aligns with Core HR, not the other way round.

Four points to watch:

• Availability of business teams, to be managed day to day.
• Data quality, especially across international entities.
• Interfaces with payroll and the ERP, which are never “plug & play”.
• The limits of the standard, to be filled with SAP-certified add-ons rather than custom developments.

Two levers make go-live safer: a phased roll-out (HR, then managers, then employees) and upskilling internal teams so they become more self-sufficient.

The principle to keep in mind

“Rigour on the standard, pragmatism on everything else.” This is the approach taken by Maïsadour, supported by Arago, and presented during our joint workshop at the convention.

Finance wants to move faster, provided everything remains traceable. Four levers to remember:

• Replan faster: according to SAP, variance alerts and scenarios suggested by Joule could make it possible to revise a plan in around 30 minutes, instead of several weeks.
• Prepare to replace SAP BPC or BFC, a priority for many customers.
• Clean up customer and supplier master data before e-invoicing. Explaining rejections in plain language also makes adoption easier for accountants.
• Include the cost of AI in every business case, and be ready to drop a use case that does not create enough value.

Another signal: SAP Concur is among the non-ERP solutions most valued by the customers surveyed by the USF. Travel and expenses remain a practical lever for controlling spend.

The message was unanimous: there is no effective AI without clean, governed and trusted data. The resulting priorities:

• Structure data as business objects (product, sales, employee, etc.) rather than tables. Slower at first, much faster afterwards.
• Ensure reversibility: avoid depending on a single AI model and keep your knowledge in-house.
• Govern your agents: traceability, compliance and control of “shadow AI” are becoming architecture topics in their own right.
• Secure the interfaces between HRIS, payroll and ERP, even within the SAP ecosystem.
• Plan SAP S/4HANA now, before resources come under pressure ahead of the 2030 deadline.

One rule applies to every project: the business owns the need, IT safeguards the consistency of the information system, and the integrator brings the expertise and prepares the run phase.