I am just back from the DSAG annual congress in Köln. The autonomous enterprise was one of the big themes, and on Thursday SAP CEO Christian Klein carried the same message in his talk on the main stage. SAP first set out the idea at SAP Sapphire in May 2026. It is a significant step for ERP. It promises faster processes, better decisions and more time for people to work on what matters. This post explains what SAP means by it, and how companies can prepare one part of the picture that every process depends on: the output it produces, as documents and as data.
What SAP means by the autonomous enterprise
SAP describes the autonomous enterprise as a model where “humans and AI work together to meet the accelerating demands of global business profitably, strategically and safely”.
It rests on three parts.
- The SAP Business AI Platform. This combines SAP Business Technology Platform, SAP Business Data Cloud and SAP Business AI. A Knowledge Graph maps business entities, processes and the relationships between them. This gives AI agents the business context they need.
- The SAP Autonomous Suite. More than 50 Joule assistants direct over 200 specialised agents across finance, supply chain, procurement and HR. Finance comes first. One example is an Autonomous Close Assistant that SAP says compresses the financial close “from weeks to days”.
- Joule Work. Instead of moving between applications and screens, a user describes the outcome they want. Agents then run the work across SAP and non-SAP systems.
SAP is rolling this out in phases, and availability varies by capability. The direction is clear, and it is a positive one. People set goals and limits. Software does more of the work to reach them. It will not only record what people do. It will decide and act.
What does not change: every process ends in output
Take a simple order-to-cash process. An agent might spot an issue, update the data, trigger the next step and move the order forward. No employee needs to handle each stage by hand.
At some point, the process reaches the outside world. Sometimes that is a document a person reads. A warehouse scans a label. A driver carries a delivery note. A customer opens an invoice. More and more often, it is structured data that another system reads. An e-invoice goes to a tax authority or an e-invoicing network. An order confirmation or a shipping notice goes straight into a customer’s or a carrier’s system.
These are the same thing in two forms. A PDF invoice and an e-invoice carry the same facts. One is read by a person, the other by a machine. Both are output, and both leave your business under your name.
Output is where a decision becomes visible to another organisation. It is also where an error becomes visible. A wrong VAT number, a missing field, a label that does not scan or an e-invoice that fails validation is not an internal problem. It reaches a customer, a carrier or an authority.
So the autonomous enterprise does not replace output. It makes good output more valuable. Agents produce more of it, faster, with fewer people looking at each document. And a growing share of it is data that no person reads at all.
What agents ask of your output
Volume goes up. When a process runs in minutes instead of days, documents and data are produced as fast as decisions are made. Peaks that once spread across a week can arrive in an hour. Output has to keep that pace without becoming the bottleneck.
Manual checks become the exception. In many companies today, someone looks at a document before it leaves. They notice the wrong logo, the missing reference or the address in the wrong field. When an agent runs the process end to end, people are freed from that work. The output then needs to be right by design. Structured data was never checked by eye in the first place. A person cannot open an e-invoice file and spot a wrong tax code. Those checks have to happen in the system.
The rules keep moving. E-invoicing formats and rules differ by country and by network, and they change often. A transaction can be correct inside the ERP while its document or data still needs the right format, structure and validation before it is sent.
Most companies run more than one system. Many businesses run SAP alongside Infor, Microsoft Dynamics or IBM i, across sites, regions or acquired companies. Agents may work across all of them. If each system produces output in its own way, the result is a different invoice layout, a different label, a different data format and a different set of rules for each one.
None of this is a reason to wait. It is a reason to get the output layer ready, so agents can deliver their full value.
Where Output Management sits
Output Management is the layer between the ERP and the outside world. It takes the data a process produces and turns it into what each recipient needs: a printed or digital document for a person, or a structured format for a system.
This is the role of InterformNG2. It takes source data from the ERP. It converts that data into the document or data format each recipient or country requires - invoices, e-invoices, delivery notes, labels and purchase orders. It validates the structure against the active rules. It then delivers the result by print, email or e-invoicing network.
This layer complements the autonomous enterprise. It sits beside it. SAP’s agents decide what should happen. Output Management makes sure the output that results, document or data, is correct, consistent and delivered in the right form.
Because InterformNG2 works with SAP, Infor, Microsoft Dynamics and IBM i, a company can manage document and data output in one place, whichever system or agent started the process. One set of templates. One set of rules. One way to see what was sent, and where.
Interform has worked with business output since 1987, and InterformNG2 runs in more than 4,500 installations worldwide. The question we help customers answer has not changed: did the right recipient receive the right document or data, in the right structure, through the right channel?
Getting your output ready for agents
If you are planning for the autonomous enterprise, your output is a practical place to start. These questions help you get it ready.
- Who checks documents today, and what do they catch? If a person corrects documents before they leave, that work needs to move into the output layer. Structured data needs the same checks, built into the system.
- How many systems produce your documents and data? If each ERP formats its own invoices, labels and data files, you have more than one set of rules to maintain.
- How do you keep up with e-invoicing changes? Know who updates formats and validation rules, and how long it takes.
- Can document rules change without changing the ERP process? If every layout change needs an ERP project, output will slow down everything around it.
- What happens when a document cannot be produced or delivered? Someone, or something, needs to know at once.
- Can you see what was sent? When an agent acts, you need a clear record of which document or data file went where, in which format, and when.
The autonomous enterprise will make decisions faster and processes simpler. The documents and data those decisions produce carry your name to customers, suppliers and authorities. With the right output layer in place, every one of them can be right, every time.
To see how InterformNG2 manages document and data output across your ERP systems, contact Interform.