Guide

SAP Business AI use cases, sorted by the strength of the evidence

As of 2026-09-25

Edited by Cédric-Olivier Mâry, sole editor and publication director — 27 years in the SAP ecosystem. How these pages are produced, and where automation is used.

SAP Business AI use cases fall into two groups that should never share a slide. The first is capability SAP documents: standard Joule agents in finance and supply chain, Joule in SAP Signavio, predictive features in SAC and IBP. The second is outcomes: figures such as Daimler Truck North America's bid-win rate rising from 10% to more than 40%, which SAP stated on its Q1 2026 earnings call and nobody has audited. Use the first to scope, and quote the second only with its source and date.

Read the umbrella before the use case

SAP Business AI is a brand, not a SKU. Since Sapphire 2026 it spans Joule as the engagement layer, the Autonomous Suite across five domains (Finance, Spend, Supply Chain, HCM, Customer Experience), seven autonomous industry solutions, and the Business AI Platform underneath. Each use case therefore has three questions attached: which layer, how it is delivered, and whether it falls under Base AI, Premium AI in AI Units, or custom-built AI on BTP runtime.

Finance and supply chain: documented agents

This is where the evidence is strongest, because SAP documents agents one by one with prerequisites. Finance includes the Accounting Accruals Agent (Restricted Availability in September 2026) and a set of closing, clearing and anomaly-detection agents listed for S/4HANA Cloud Private Edition. Manufacturing includes the Production Planning and Operations Agent, which releases up to 20 production orders per run and pauses for planner confirmation.

Our MRP exception pattern for discrete manufacturers — triage, cluster by root cause, escalate only what threatens delivery — is an architecture pattern built on these pieces. Its time-saving figures come from pilots our card cannot trace to an audited source, so treat them as a hypothesis to test.

Process and architecture work: Signavio and LeanIX

Joule with SAP Signavio solutions became generally available on 10 February 2026, covering informational, navigational and transactional requests on process content. On the architecture side, SAP LeanIX's AI features arrived in waves through Q1, Q2 and September 2026. Both are capability statements with SAP release dates, not outcome claims — the right material for a scoping workshop.

Industry use cases: where regulation shapes the design

Banking splits into AML alert scoring, KYC dossier synthesis and credit scoring; under the EU AI Act, credit scoring and AML fall in the high-risk class, with stand-alone high-risk obligations applying from 2 December 2027 after the Digital Omnibus. Pharma clinical-trial supply adds GxP traceability for every AI proposal. Retail and CPG must separate demand sensing from dynamic pricing, which vendors often bundle. Public-sector AI in Europe is scoped to what can legally run inside a sovereign boundary, such as SAP on Bleu in France or Delos Cloud in Germany.

These cards describe patterns and constraints. None of them is an SAP-published customer result.

Vendor-cited outcomes, and how to reuse them

SAP's Q1 2026 call cited Daimler Truck North America (bid wins from 10% to above 40%, EUR 70 million impact within 12 months), EY (up to 30%), KPMG (up to 20% faster sprints) and Bosch Digital (plus 20% developer productivity). The DTNA statement names SAP Business AI, not Joule, and gives no baseline or method. Our reuse rule: cite as SAP-reported with the date, never transfer the percentage to another client, and measure the client's own baseline.

Turning a use case into a business case

A use case becomes defensible when build, run and governance costs sit on one quarterly tracker next to a named owner, a baseline and a measurement cadence. Benefits accrue slowly and partially while costs are upfront; a benefit without a baseline can only be claimed, never realised.

What we cannot assert

The EY, KPMG, Bosch Digital and Daimler Truck figures are SAP-cited on an investor call, not independently audited. Percentage gains in our industry pattern cards (MRP exceptions, AML scoring, credit scoring) are not traced to an audited SAP customer result, so we do not repeat them here.

Frequently asked

Which SAP Business AI use cases are generally available?

Check per agent and per edition. Examples with SAP-documented status in our corpus: Joule with SAP Signavio (GA 10 February 2026), the agent builder in Joule Studio (GA per Q4 2025 highlights), and the Accounting Accruals Agent (Restricted Availability in September 2026, GA planned Q1 2027).

Is the Daimler Truck 10% to 40% bid-win figure audited?

No. It was stated by SAP on its Q1 2026 earnings call on 23 April 2026. It names SAP Business AI without a product list, baseline definition or method for the EUR 70 million impact.

Which SAP AI use cases are high-risk under the EU AI Act?

Our banking card classes AML transaction scoring and credit scoring as high-risk under Article 6 and Annex III. Stand-alone high-risk obligations apply from 2 December 2027 under the Digital Omnibus, in force since 27 July 2026.

How do I measure ROI on an SAP AI use case?

Set build, run and governance costs against a baseline per use case before go-live, with an owner and a quarterly cadence. Labour hours saved are easiest to measure; cycle time and decision quality matter more.