Now accepting guest editorial from SAP, Oracle, Workday and ServiceNow practitioners.Pitch a guide
The Sign-Off Table
All statistics

Statistics · Published

Test Automation Statistics (2026): Coverage, AI and Quality Data for SAP and ERP Teams

60% of organizations say they deploy untested code to production, according to Tricentis' 2026 Quality Transformation Report.

Compiled by Matt Angerer
24 sources · 12 min read

Automation coverage, AI adoption and release quality are the three numbers most ERP test leads get asked about. This roundup collects the 2025 and 2026 figures we could source for those topics, plus SAP-specific migration data that frames the testing workload. Where two credible sources disagree, we show both.

Key takeaways

67%

of SAP users surveyed had fully or partly migrated to S/4HANA (2026 survey of SAP users)

Source: Precisely and ASUG, 2026

60%

of organizations deploy untested code to production (2026 Tricentis survey, 2,501 IT and QA leaders)

Source: Tricentis, June 2026

57%

of QA tests are automated on average (2026 Sembi Software Quality Pulse, nearly 4,000 respondents)

Source: Sembi Software Quality Pulse Report, 2026 (reported by Ranorex)

9 months

median ERP project timeline (Panorama 2026 ERP Report, 170 organizations, January 2025 to January 2026)

Source: Panorama Consulting Group, 2026

S/4HANA migration status: where the workload comes from

Migration volume sets the testing load for every project. The most-cited survey of SAP users shows steady growth, but the definitions and samples differ across publishers.

Source: Precisely and ASUG, 2026

  • The 2026 Precisely and ASUG report puts the migrated share at 67%, with another 23% expecting to finish within 24 months. Those figures were 59% in 2025 and 50% in 2024 (Precisely and ASUG, 2026).
  • SAPinsider's 2025 benchmark found 27% of organizations in active implementation and 21% still in evaluation (SAPinsider, 2025).
  • The same SAPinsider report put 32% of respondents as already transitioned. Its detailed findings PDF gives 31% for the same question, a minor discrepancy we flag rather than resolve (SAPinsider benchmark detailed findings, February 2025).
  • Gartner found in March 2025 that 61% of SAP customers on ECC had not yet licensed S/4HANA (The Register, citing Gartner, October 2025).
  • An aggregated estimate from early 2025 put completed ECC-to-S/4HANA moves at 28% to 33%. The source itself warns the studies used different samples and methods (Datalink Dynamics, 2025).

The 2026 Precisely and ASUG figure is a survey of SAP users who respond to ASUG research. It is not a census of all ECC customers. A separate Quantumrun compilation reports that more than 60% of legacy SAP ECC customers have not migrated (Quantumrun, SAP statistics 2026, secondary compilation). That number is not directly comparable with the ASUG figure, but both point to a large remaining backlog.

67%
61%

Source: Precisely and ASUG, 2026; Gartner via The Register, 2025

How much of testing is automated

Two 2026 surveys give very different automation levels. The gap is largely a definition problem. One measures the share of QA tests automated. The other measures average automation coverage.

QA tests automated (Sembi, 2026)57%
Average automation coverage (WQR 2025-26)33%

Source: Sembi Software Quality Pulse Report, 2026 (via Ranorex); World Quality Report 2025-26, as summarized by testresults.io

  • The Sembi report found teams had automated 57% of their tests on average, yet nearly half remain neutral or dissatisfied with their QA process (Ranorex summary of the Sembi report, 2026).
  • The Capgemini-led World Quality Report 2025-26 reports average automation coverage of 33%. Nearly half of organizations were still in planning, and only 8% described a fully established automation strategy (testresults.io summary of WQR 2025-26).
  • The World Quality Report 2024-25 edition reported a global average test automation level of 44%, as quoted by a reviewer. We have not verified that figure against the primary report, so it is not used as a trend point (Reviewer quoting WQR 2024-25).
  • In ASUG-linked SAP research, automation adoption was reported to have plateaued at 57% while migration absorbs resources. That is a separate 2025 measure from the Sembi figure (Precisely and ASUG survey coverage, 2025).

Historical (more than five years old): An earlier Tricentis and ASUG survey found that 41% of respondents saw a lack of automated testing as a concern during S/4HANA work. It also found that 76% expected test automation to become more important (CIO, summarizing Tricentis and ASUG research; Tricentis, SAP testing expectations). Use these as context only.

Release quality and the cost of skipping testing

The 2026 Tricentis report measures the share of organizations that ship code without full testing, a common sign-off risk.

60%

Source: Tricentis, June 2026

Focus on delivery speed (2025)45%
Focus on software quality (2025)13%

Source: Tricentis, May 2025

Timelines and budgets in ERP programs

Panorama Consulting's 2026 report is a general ERP benchmark, not an SAP-only study. Its figures frame how much schedule slack a test plan actually has.

  • The Panorama 2026 ERP Report analyzed 170 organizations with data collected between January 2025 and January 2026. Its median project timeline was nine months (Panorama Consulting Group, 2026).
  • A summary of the same report says 30% of projects exceeded budget, with 22.9% over slightly and 7.1% over significantly (Xorosoft, summarizing Panorama 2026).
  • The same summary reports 22.3% of projects finished late (Xorosoft, summarizing Panorama 2026).
  • In a Freeform Dynamics survey of 455 respondents, 95% said building a positive case to migrate is a big effort or genuinely challenging. Rimini Street sponsored that research, so weigh it accordingly (The Register, October 2025).
Projects over budget30.0%
Projects completed late22.3%

Source: Xorosoft summary of Panorama Consulting, 2026

AI in testing: adoption is broad, gains are narrow

AI claims are the most inflated category of testing data. The figures below separate use from measurable results.

Use AI somewhere in testing (BrowserStack, 2026)94%
Use AI across most testing workflows (BrowserStack, 2026)61%
Still testing manually despite automation (BrowserStack, 2026)92%

Source: BrowserStack 2026 leadership survey, as reported by Quash

  • The BrowserStack 2026 leadership survey, a vendor-published study, reports 94% using AI in testing and 61% using it across most workflows. It also found 92% still testing manually, though 87% had automation in place (Quash, reporting BrowserStack 2026).
  • The Sembi report found that only 17% of teams say AI-driven testing has delivered significant gains. Meanwhile 61% report higher testing demand from AI-generated code (Ranorex summary of Sembi 2026).
  • In WQR 2025-26, 15% of organizations had scaled Gen AI across quality engineering. Another 30% used it in operational testing workflows. Among adopters, average productivity gains were 19% (testresults.io summary of WQR 2025-26).
  • Gartner predicts that by 2030 AI agents will start, coordinate, run and analyze more than 50% of enterprise software testing. That figure is quoted here from a vendor's write-up of the Gartner research (TestMu AI, quoting Gartner, 2026).
15%

Source: World Quality Report 2025-26, as summarized by testresults.io

Test data: the constraint behind many automation delays

Capgemini's WQR 2025-26 places test data near the top of the barriers list, and it tracks synthetic data use over two years.

Synthetic data use (WQR 2024)14%
Synthetic data use (WQR 2025)25%

Source: Capgemini, World Quality Report 2025-26

  • Capgemini reports that synthetic data use rose from 14% in 2024 to an average of 25% in 2025 (Capgemini, WQR 2025-26).
  • The same report found 60% of organizations struggle to provide secure, scalable test data. Another 58% cite difficulty adopting AI-powered tools (Capgemini, WQR 2025-26).
  • Among those surveyed, 95% use AI for test data generation, but only 10% have fully integrated AI-driven test data management (testresults.io summary of WQR 2025-26).

Skills, tooling and governance

  • Tricentis' 2026 report says 53% of teams manage six to 10 AI or automation tools. Only 35% feel fully prepared to govern AI agents at scale (Tricentis, 2026 Quality Transformation Report).
  • The same report notes trust in AI agents to make release decisions fell from 48% to 34% year over year (Tricentis, 2026).
  • Among ASUG-linked respondents, 33% had adopted a best-of-breed automation approach, up from 19% in 2025. 84% call no-code or low-code capability important when choosing a platform (Precisely and ASUG, 2026).
Feel somewhat prepared to govern AI agents (Tricentis, 2026)82%
Feel fully prepared to govern AI agents at scale (Tricentis, 2026)35%

Source: Tricentis social post summarizing its 2026 report

What this means for your sign-off

  • Set the automation target by definition. A 57% share of tests automated and 33% average coverage can both be accurate. Put the measure you will report in the entry criteria before the first cycle starts.
  • Treat untested change as a go/no-go item. When 60% of organizations report deploying untested code, your release evidence needs to show what was actually executed. Link your sign-off pack to the regression results, not to a count of scripts.
  • Budget for test data first. With 60% struggling on secure, scalable test data, plan masking, copies and synthetic sets early. See test data for S/4HANA for the options.
  • Automate the repeating regression, not everything. The migration backlog is large, and your automation should follow repeated business processes. See S/4HANA regression automation: what to automate first.
  • Plan against a nine-month median, not a fixed deadline. Panorama's median gives a useful starting point. Your own N+1 landscape and cutover windows will set the real schedule. See SAP N+1 landscape testing.
  • Measure AI output before you trust it. Use the free template library for an automation ROI calculator and an AI-generated test review checklist before you adopt AI-generated scripts into a sign-off set.

Frequently asked questions

What share of SAP customers have migrated to S/4HANA?

In a 2026 Precisely and ASUG survey of SAP users, 67% reported being fully or partly migrated. Gartner's March 2025 finding, reported by The Register, is different: 61% of SAP ECC customers had not yet licensed S/4HANA.

How much of QA testing is automated in 2026?

The 2026 Sembi Software Quality Pulse report puts the share of QA tests automated at 57%. The World Quality Report 2025-26 reports average automation coverage of 33%. The two measures are not the same.

How many organizations deploy untested code?

Tricentis' 2026 Quality Transformation Report found that 60% of organizations ship code to production without full testing, based on a survey of 2,501 IT and QA leaders across six countries.

How long does an ERP implementation take?

The Panorama Consulting 2026 ERP Report, covering 170 organizations, reports a nine-month median project timeline. Its sample is general ERP, not SAP-only.

How many organizations use AI in testing?

A BrowserStack 2026 leadership survey, a vendor-published study, found 94% use AI somewhere in testing and 61% across most workflows. The World Quality Report 2025-26 found that 15% had scaled Gen AI across quality engineering.

How much test data is synthetic?

The World Quality Report 2025-26 reports that synthetic data use rose from 14% in 2024 to an average of 25% in 2025.

Will AI agents run most enterprise software testing?

Gartner's forecast, as quoted by TestMu AI, is that by 2030 AI agents will initiate, coordinate, execute and analyze more than 50% of enterprise software testing. Gartner's May 2026 research argues that agentic testing platforms are needed to manage nondeterministic AI behavior.

Sources

  1. Precisely and ASUG survey finds 67% of SAP users have fully or partly migrated to S/4HANA, Precisely and ASUG, 2026
  2. Tricentis Report: 60% of Global Organizations are Shipping Untested Code, Tricentis, June 2026
  3. 2026 Quality Transformation Report, Tricentis, 2026
  4. Two-Thirds of Global Organizations Significantly at Risk of Software Outage, Tricentis, May 2025
  5. The State of Test Automation in 2026: Key Findings from the Software Quality Pulse Report, Sembi via Ranorex, 2026
  6. World Quality Report 2025-26, Capgemini, 2025
  7. World Quality Report 2026: is QE now a leadership issue?, testresults.io summary of WQR 2025-26, 2025
  8. Reviewing Capgemini's World Quality Report 2024-25, reviewer summary, October 2024
  9. The 2026 ERP Report, Panorama Consulting Group, March 2026
  10. ERP integration failure statistics 2026, Xorosoft summary of Panorama, 2026
  11. SAP S/4HANA Migration 2025, SAPinsider, 2025
  12. SAPinsider Benchmark Report 2025: Detailed Findings, SAPinsider, February 2025
  13. Precisely Survey: SAP S/4HANA Migration Pace Accelerates, Precisely and ASUG, 2025
  14. SAP users still wrestling with business case for S/4HANA, The Register, October 2025
  15. How Many SAP Customers Have Migrated to S/4HANA, Datalink Dynamics, 2025
  16. SAP Statistics 2026, Quantumrun, 2026 (secondary compilation)
  17. SAP Trends to Watch: SAP S/4HANA, Automation, and Anticipating Risk, CIO (historical)
  18. SAP S/4HANA, Agile, Cloud, and Testing Transformations, Tricentis (historical)
  19. Freeform Dynamics survey coverage (sponsored by Rimini Street), The Register, October 2025
  20. AI Testing Statistics: Insights from World Quality Report, Quash, citing BrowserStack 2026, 2026 (secondary)
  21. TestMu AI named a Leader in the 2026 Gartner Magic Quadrant, TestMu AI quoting Gartner, 2026
  22. AI-Driven Software Engineering Requires Agentic Testing Platforms, Gartner, May 2026
  23. Tricentis Research: Manufacturers Embrace AI, Manufacturing Outlook, 2026
  24. Tricentis Report Examines Software Quality as AI Use Expands, ExecutiveBiz, 2026

Last updated: November 6, 2026. Statistics are refreshed as new research is published.

More statistics

How this page is made: statistics are gathered from published research with AI-assisted search, every figure is linked to its source, and the page is refreshed as new data is published. Found an error? Tell us.