AI IN SAP S/4 HANA P2P

AI in the S/4HANA Procure to Pay process

Victor van den Hazelkamp - Managing Partner

AI in the S/4HANA Procure to Pay process: a dream for the future or already tangible value?

AI is everywhere. No SAP event, blog, or roadmap goes by without artificial intelligence. It's high time to explore what's already available within the Procure to Pay (P2P) process and what future developments look like.

This blog provides an overview of three AI solutions that SAP already uses in the P2P process: SAP Document AI, SAP Joule, and SAP Ariba Central Invoice Management. It then discusses the SAP roadmap, focusing on the further development of these AI solutions. Finally, the blog highlights use cases for organizations that want to go beyond standard functionality. The focus is on S/4HANA, both the private and public cloud versions.

SAP's current AI solutions within the P2P process

1. SAP Document AI – Intelligent document processing

What is it?

SAP Document AI is intelligent document processing that helps organizations process large numbers of business documents with varying content and structures, such as invoices and purchase orders. This SAP solution extracts and enriches data from these files, regardless of how the information is organized in the document. This data is then transferred to target systems, such as SAP S/4HANA.

What makes this solution powerful?

• Multichannel. Documents are processed through various channels, such as central email inboxes and cloud storage, or through ad hoc document uploads in an intuitive user interface or SAP Mobile Start.

• User-centric UX. An intuitive user interface is designed for efficient and easy validation of incoming documents.

• File processing & OCR (Optical Character Recognition). Thanks to advanced OCR technology, a large number of business documents (including invoices, purchasing, and HR documents) are supported in various formats and in more than 100 languages.

• Preconfigured content. SAP Document AI offers a large number of pre-trained models (known as schemas) for standard documents such as invoices and purchase orders. Its strength lies in its ability to create custom schemas for every document type required within the organization.

View fullsize

Figure 1: Business documents supported by SAP Document AI

Figure 2: Reliability percentages of extracted fields

• Generative AI & Instant Learning (Premium edition only). For both standard and custom schemas, SAP Document AI uses generative AI and Large Language Models (LLMs) to understand and extract information. Generative AI is applied to assign reliability percentages to the extracted data. These percentages can be used in the further processing of the business document within S/4HANA.

  With Instant Learning, the model learns directly from manual corrections. After one correction, the system automatically recognizes the same document layout correctly during subsequent processing.

2. SAP Joule – AI Co-pilot within P2P

What is it?

SAP Joule is SAP's generative AI co-pilot that supports users in working in SAP through natural language, context awareness, and built-in AI agents.

SAP Joule is an AI assistant that:

• Understands what a user is doing in SAP (context).

• Answers questions in natural language.

• Summarizes, explains, and analyzes information.

• Helps with navigation and faster task execution.

What are the key features of SAP Joule?

• Navigation. Joule helps users quickly find the right apps and features without needing to know what each app or feature is intended for. 

  Example question from user: "How can I view supplier invoices?"

  SAP Joule automatically navigates to the Manage Invoicing Documents app.

• Informative. Joule answers questions about SAP functionality. This can be based on SAP Help (S/4HANA private cloud edition) and/or on our own documentation stored on SharePoint (S/4HANA public cloud edition), for example.

•  Transactional. Joule provides users with insight into business data. This can include documents such as contracts, purchase orders invoices, as well as master data such as business partners, general ledger accounts, and cost centers. Joule also supports users in creating business data and accelerating task completion.

  Sample questions from users:

  • purchase orders all purchase orders from supplier X.

  • Show all open purchase orders approver Y.

  • Change the delivery date of purchase order to date Y.

•  Analyses. Joule helps users obtain insights to support decision-making. A question in natural language results in a graph, table, or other visual overview. SAP Analytics Cloud (SAC) is required for this functionality.

   Example question from user:

  Show the top 5 suppliers based on purchase value.

3. SAP Ariba Central Invoice Management

What is it?

SAP Ariba Central Invoice Management (CIM) is a solution on SAP Business Technology Platform (SAP BTP) for the central receipt and processing of purchase invoices. SAP CIM integrates with various backend systems, including SAP public cloud edition. More information on this topic can be found in the blog by our colleague Tjalling Jan Gerkema.

AI functionalities

SAP CIM has AI functionalities that automatically enrich non-PO invoices. This applies to fields such as general ledger account, cost center, and WBS (Work Breakdown Structure) element. This makes the processing of these invoices more efficient, simpler, and less prone to errors.

The SAP roadmap: future AI developments in P2P

Looking at the SAP roadmap for 2026, it shows a clear direction for each of the AI solutions described.

1. SAP Document AI

Developments within SAP Document AI are happening at a rapid pace. New functionalities are also on the agenda for 2026. Here are a few examples:

• Expansion of support for customized Large Language Models (LLM).

• Reliability percentages at document level instead of field level. This makes it easier to determine which documents still need to be validated and which can be processed immediately.

• Translation of documents into your own language, making validation easier.

• Improved monitoring of the various import channels, such as mailboxes. Clear insight into the import status (success/error) enables immediate action without having to study the technical logs. 

2. SAP Joule

SAP Joule is an important spearhead for the further development of generative AI within SAP. Various developments within the P2P process are planned for 2026. Some examples for the public cloud edition are:

• Identification of long-outstanding approval tasks for purchase orders. Joule can generate automatic email reminders to encourage action.

• Generating an overview of purchasing contracts, including insight into the validity period, approval status, and contract utilization. 

3. SAP Ariba Central Invoice Management (CIM)

The following AI innovations are planned for SAP CIM in 2026:

• Integration with SAP Document AI, enabling CIM to utilize generative AI and Instant Learning.

• Manual creation of invoices via Joule.

• Invoice Exception Handling Agent. An AI assistant that automatically detects invoice exceptions and either resolves them immediately or proposes a solution.

• Invoice Dispute Management Agent. An AI assistant that helps automate the handling of disputes with suppliers, including detection, determining follow-up action (e.g., rejecting or requesting a credit invoice), collecting necessary documents, and communicating with the supplier.

Use Cases: Additional AI Applications in P2P

Within the P2P process, SAP's AI innovation focus is currently on SAP Document AI, SAP Joule, SAP CIM, and the SAP public cloud edition. Standard AI agents and optimizations for the P2P process within S/4HANA private cloud are not currently explicitly included in the SAP roadmap.

However, SAP offers various tools for developing AI applications yourself and integrating them into the P2P process.

Some possible use cases are:

• Proposal for a supplier when creating an ATB/PO (Technical purchase order Request purchase order Order).

• Proposal for classification when creating an ATB/PO.

• Automatic selection of the correct contract when creating an ATB/PO.

• Compliance check after entering a PO.

• Explanation for a price or quantity block on an invoice, including the proposal for and implementation of concrete actions (such as changing a PO or entering a goods receipt/SES).

• Improvement of invoice matching.

• Automatic determination of the correct approver for a non-PO invoice.

Conclusion: Let's get started!

SAP offers concrete AI solutions within the P2P process that can add immediate value for organizations. SAP continues to innovate in this area and offers various tools for developing AI solutions and integrating them into the P2P process, both within the public and private cloud editions of S/4HANA.

In short, development is in full swing. Seize the moment and get started with AI to further optimize the P2P process.

Learn more

Please feel free to contact Victor van den Hazelkamp. He will be happy to discuss the use and application possibilities of AI within the SAP P2P process.