SAP Business AI and Joule Adoption Statistics (2026)

Q: How many SAP customers run SAP Business AI in production?

In the DSAG Investment Report 2026, secondary coverage reports that just 3 percent of AI-in-production respondents ran an SAP-based AI use case. The survey covers DACH-region DSAG members, so it is not a global count.

Q: How many AI agents and skills does SAP ship?

Secondary coverage cites over 40 specialized AI agents and more than 2,400 Joule Skills in Q1 2026. Forrester's count of agents unveiled at Sapphire 2026 is higher, so check which definition a source uses.

Q: What share of SAP's large deals include AI?

SAP's Q3 2026 call said 90% of the 50 largest deals included Business AI and Business Data Cloud.

Q: How many RISE customers get Joule assistants?

SAP's Sapphire announcement said RISE with SAP customers will have three assistants activated within their first year.

Q: How many enterprises have moved AI pilots into production?

Deloitte's 2026 survey found that only 25% of respondents have moved 40% or more of their AI pilots into production.

Q: What does Gartner forecast for agentic AI project cancellations?

Gartner's forecast is that over 40% of agentic AI projects will be canceled by the end of 2027. This is a forecast, not an observed rate.

SAP now presents Business AI and Joule as the main reason customers move to the cloud, and its quarterly numbers show AI-linked deal volume. What customers actually run in production is a separate question, and the sources disagree on several figures. This roundup keeps SAP-reported numbers, customer surveys and analyst forecasts apart so you can see which is which.

Key takeaways

90%

of SAP's 50 largest Q3 2026 deals included Business AI and Business Data Cloud

Source: SAP Q3 2026 earnings call, transcript, Oct 2026

43%

of DSAG-surveyed SAP user companies had implemented at least one AI use case, including pilots

Source: heise, citing DSAG Investment Report 2026, Oct 2026

77%

of DSAG respondents with AI in production used non-SAP tools (secondary coverage)

Source: IgniteSAP, 2026

40%

of enterprise applications forecast by Gartner to include task-specific AI agents by end-2026

Source: Gartner, Aug 2025

Adoption signals from SAP's own reporting

SAP's figures are the most current and cover a broad installed base, but they measure contracts, revenue and platform reach rather than how many users work with Joule each week. Treat them as vendor data.

In Q3 2026, total revenue reached EUR 10.3 billion, up 12% year over year, and cloud revenue hit EUR 6.7 billion, up 25%.
The cloud backlog stood at EUR 24.1 billion. Growth is reported inconsistently: one source puts it at up 26%, while another says it climbed 28% to €24,144 million. See the conflicts section below.
The prior quarter, Q2 2026, showed a current cloud backlog of €22.9 billion, up 27%, with cloud ERP Suite revenue up 25%.
Within the quarter, Cloud ERP Suite revenue jumped 28% and now represents 89% of total cloud revenue, while software licenses fell 9% and services revenue declined 5%. SAP attributes the services decline to a deliberate shift of billable hours toward adoption support for AI solutions.
Cloud gross margin moved the other way: it reached 74.3% on a non-IFRS basis, down 0.7 percentage points year over year, which SAP ties to AI investment and higher infrastructure costs.

90%

Source: SAP Q3 2026 earnings call transcript, Oct 2026. Note: the transcript says nine of 50; a secondary summary says nine of ten (see conflicts).

On the product side, SAP's own Q1 2026 release notes say Joule is now live across 35 solutions. SAP also said at Sapphire that RISE with SAP customers will have three assistants activated within their first year, while SAP GROW customers receive full portfolio access at onboarding.

Agent and skill counts differ by source. One analyst-facing summary reports over 40 specialized AI agents, more than 2,400 Joule Skills, and an open agent-to-agent protocol across 35 SAP solutions in Q1 2026.
Forrester commentary on Sapphire 2026 counted 224 agents and 51 assistants, and noted that most sat in general availability, Early Adopter Care or preview status.
SAP's Sapphire keynote described hundreds of out-of-the-box agents cutting across all core business processes, which is a broader definition than the 40+ count.
A secondary market roundup reports that Joule adoption rose ninefold through 2025. We did not find SAP's primary release defining the metric, so treat this as unverified.

What customers are actually running (DSAG)

The DSAG Investment Report 2026 is the most-cited independent customer survey in this roundup, and it has an important scope limit: it covers the German-speaking user base.

The survey collected 198 responses between December 8, 2025 and January 21, 2026.
The sample is limited to companies surveyed in Germany, Austria and Switzerland (DACH), so it does not have a global sample.
In the January survey, 43 percent had implemented AI use cases, including many tests and pilot projects.
Among firms with AI in production, the reported split is stark: 77 percent use non-SAP solutions, while only three percent rely on SAP's own AI tools. Secondary coverage describes the three percent as SAP customers running SAP Business AI in production, but the denominator is not consistent across write-ups.
Interest is higher than use. 45 percent of DSAG members give AI significant consideration in their IT budgets, and 48 percent see high potential benefits.
Deployment model matters for timing. 56 percent of respondents use S/4HANA on premises, and 42 percent plan medium or high investments in this operating model this year.

43%

56%

Source: heise, citing DSAG Investment Report 2026, Oct 2026

Non-SAP tools (AI in production)77%

SAP-based AI (AI in production)3%

Source: IgniteSAP, 2026, summarizing DSAG Investment Report 2026

Give AI significant consideration in IT budgets45%

See high potential benefits48%

Implemented an AI use case43%

Source: e3mag, 2026, on DSAG Investment Report 2026

Analyst forecasts for agents in enterprise applications (Gartner)

Gartner's numbers are forecasts and surveys, not measured market totals. Read them as direction and risk, not as a count of installed agents.

Gartner predicted that 40% of enterprise applications will be integrated with task-specific AI agents by the end of 2026, up from less than 5% today.
On the risk side, Gartner forecast that over 40% of agentic AI projects will be canceled by the end of 2027, due to escalating costs, unclear business value or inadequate risk controls.
That forecast rests on a small poll. In January 2025, a Gartner poll of 3,412 webinar attendees found 19% said their organization had made significant investments in agentic AI, 42% had made conservative investments, 8% no investments, and 31% were waiting or unsure.
Gartner also estimated that only about 130 of the thousands of agentic AI vendors are real.

Significant agentic AI investment (Jan 2025 poll)19%

Conservative investment (Jan 2025 poll)42%

No investment (Jan 2025 poll)8%

Wait-and-see or unsure (Jan 2025 poll)31%

Source: Gartner, June 2025 (poll conducted January 2025, historical)

Enterprise apps with task-specific agents, 2025 (less than)<5%

Enterprise apps with task-specific agents, end-2026 (forecast)40%

Source: Gartner, Aug 2025

Enterprise-wide agentic AI (Deloitte and KPMG)

Deloitte's 2026 State of AI in the Enterprise is the broadest survey in this list outside SAP's own reporting. It spans many industries, so it is not SAP-specific.

Deloitte surveyed 3,235 information technology and business leaders from 24 countries, all with direct involvement in their organizations' AI programs.
Pilot conversion is the bottleneck. Only 25% of respondents have moved 40% or more of their AI pilots into production.
Planned agent use is high. By 2027, 74% of respondents expect their companies to be using AI agents at least moderately.
Governance lags. Only 21% of respondents say their organizations have a mature governance model in place for agentic AI.
Among firms reporting results, 66% report efficiency and productivity gains, 53% report better decision-making and data-driven insights, and 20% report revenue growth from their AI initiatives.

Expect at least moderate agent use by 202774%

Moved 40%+ of AI pilots into production25%

Mature agentic AI governance model21%

Source: Deloitte, 2026 State of AI in the Enterprise (survey Aug to Sep 2025); pilot figure from Deloitte press room, Jan 2026

Efficiency and productivity gains66%

Enhanced decision-making53%

Revenue growth from AI20%

Source: Deloitte, AI agents for business, 2026

A secondary market roundup cites a KPMG Q3 2026 US figure, but we did not confirm it against KPMG's publication, so we leave it out of the counts above.

Cost and business case

SAP reports that its own internal AI use is paying off. Management said internal AI adoption among its 110,000 employees produced 20% productivity gains in corporate functions and a 70% increase in developer code contributions. These are management-reported figures, not independently audited.
SAP is also funding partner adoption. It launched a €100 million fund for partners to help customers deploy SAP-built AI assistants and agents.
Gartner's cancellation forecast names cost as a leading cause, which lines up with the margin pressure SAP reports above.

Timelines: the ECC clock and the commitments in RISE and GROW

AI commitments now sit inside migration contracts, which changes how you should time testing. The clock that matters most is the ECC end of regular maintenance.

In the DSAG survey context, regular maintenance of SAP ECC ends at the end of 2027. Our ECC end-of-maintenance explainer covers the later dates.
Migrations to S/4HANA require money and personnel, which DSAG sees as a constraint on AI spending as well.
SAP is offering a no-cost on-ramp. Through the end of 2026, SAP customers and partners can receive free design-time access to Joule Studio, including AI-assisted development under fair-use limits.
Joule Work is rolling out to the customer base. SAP said all SAP customers will enjoy the benefits of Joule Work when it rolls out later this month (October 2026 announcement).

Risk, governance and testing

We did not find a sourced, SAP-specific statistic on how many AI-generated test cases or agent-driven changes reach production without defects. We would rather leave that gap visible than fill it with an estimate. The sourced risk data points in one direction: governance and value proof, not model capability.

DSAG said that for AI to be enterprise-ready in SAP, it needs orchestrable agents, transparent decision-making processes, secure data foundations, and open modularity that securely integrates third-party agents.
DSAG's own list of success factors adds data quality, integration effort, governance, and comprehensible costs.
SAP's CEO argued at Sapphire that context is the main reason enterprise AI projects fail to deliver value. That is a vendor claim, not a measured rate.
Gartner's cancellation drivers, cost, value and risk controls, map directly onto the test-plan questions your sign-off should answer.

For the practical side, our template library includes an AI Agent Guardrails and Approval Gate Register and an AI-Generated Test Review Checklist.

Where the numbers conflict

Cloud backlog growth. One source reports up 26%; another reports a 28% climb to €24.1 billion. We show both and did not pick one.
Top-deal share. The earnings transcript says 90% of the 50 largest deals included Business AI and Business Data Cloud. A secondary summary says nine of SAP's ten largest deals. The figures are different bases, so the 90% number is the one tied to SAP's own wording.
Agent counts. The sources give 40+ agents, 224 agents, and "hundreds" of agents, depending on whether they count generally available, early-adopter and preview offerings.
DSAG denominators. The 3% and 77% figures come from secondary write-ups that describe the base differently. DSAG's own press material says the member companies have put few AI use cases into productive use but does not give the percentage in that release.
DSAG sample size. DSAG says significantly fewer large companies were surveyed than in the last survey, while an independent analysis says more large enterprises are reflected. We flag the disagreement and do not use the sample composition to weight any figure.

What this means for your sign-off

Ask which AI is in scope. If the 77% versus 3% split holds in your landscape, you may be testing non-SAP AI tools that touch SAP data. Your scope and interface tests need to cover both.
Test the vendor commitment, not just the feature. RISE customers get three Joule Assistants activated within the first year. Write acceptance criteria for activation, not only for the functions those assistants touch.
Put governance in the exit criteria. With only 21% of firms reporting mature agentic governance in Deloitte's survey, a sign-off that does not record approval gates and rollback paths is weak.
Measure before you scale. Gartner's cancellation forecast points to unclear value as a driver. Set a baseline in the pilot, as our SAP AI for practitioners guide describes.
Use Joule with a test plan. If your team uses Joule for test design or query work, our Joule prompts guide for S/4HANA projects covers the practical patterns.
Plan around the ECC date. With regular ECC maintenance ending at the end of 2027 for the DSAG population, the AI decision sits inside the migration timeline, not beside it.

Frequently asked questions

How many SAP customers run SAP Business AI in production?