Skip to content
digitice.io
AI-first How we work Cases AI Radar Team
DEEN
Book a conversation ↗
digitice.io / AI Radar / All studies

AI RADAR · STUDY · 20 Oct 2026

Value and scaling

Few respondents report rules for stopping AI initiatives

The survey distinguishes rules before production from rules for AI already in use. It measures reported practices, not proven business success.

The survey distinguishes rules before production from rules for AI already in use. It measures reported practices, not proven business success.

Source

AI initiatives need stopping rules

Gartner · Gartner · 19 Oct 2026

Source finding

31% of surveyed CIOs and technology executives report guidelines for deciding when AI pilots should not enter production. 17% report guidelines for abandoning low-value AI already in production. These are separate practices, not parts of a distribution.

Our interpretation

Businesses should make explicit decisions on continuing and stopping live applications as well as launching them. The survey does not establish that these rules alone cause better business performance.

Method and evidence

Gartner surveyed 2,002 CIOs and technology executives from May to July 2026. Moderate evidence for reported practices, limited evidence for effects. The public summary does not fully disclose the questionnaire, weighting or industry distribution.

Limitations and interests

Self-reported practices, not an independent audit of AI value. Associations do not establish causation. Gartner sells research and advice. The publicly checked release is not the full client report.

What businesses can do

Set minimum benefits, full costs, review dates and accountable owners for pilots and live applications. Use these criteria to decide whether to continue, change or stop.

Where to start

Read the survey description and the paragraphs about stopping rules before production and during live operations. Published on 19 October 2026; 2027 identifies the report edition.

Open original study ↗

← Previous studyOrganised team chats: more correct AI answers in the test
Next study →At Anthropic, more Claude sessions focus on instructions and oversight

More from this area: Value and scaling

  • 25 Oct 202683% report no productivity impact from AI yet
  • 24 Oct 2026Using AI does not mean embedding it in core operations
  • 23 Oct 2026More speed with AI does not remove every bottleneck

All studies in the AI Radar → · ← Back to the AI Radar

digitice.io

Building AI-first companies.

FAQImprintPrivacy
© 2026 digitice GbR