Notifications
Clear all
Topic starter
04/07/2026 10:00 pm
TL;DR: Automation fails when teams layer it on undocumented, measurement-blind workflows, according to JumpCloud and cited research from S&P Global Market Intelligence, Ernst & Young, Gartner, and McKinsey. The practical lesson is that process standardisation, measurement, and control must come before workflow automation, or inefficiency and exceptions scale instead of shrinking.
NHIMG editorial — based on content published by JumpCloud: Automation Mindset: why broken processes break automation
By the numbers:
- Automation project abandonment rates jumped from 17% to 42% in a single year.
- 30 to 50% of RPA scripts break immediately after deployment.
- 60% of AI initiatives will be abandoned through 2026 due to untrustworthy outputs.
Questions worth separating out
Q: What breaks when teams automate an undocumented workflow?
A: Undocumented workflows break because automation removes the human judgement that was compensating for missing inputs, unclear handoffs, and inconsistent approvals.
Q: Why should process standardisation come before automation in IAM?
A: Process standardisation comes first because automation scales whatever exists, including ambiguity.
Q: How do security teams know a workflow is ready for automation?
A: A workflow is ready when the team can define it clearly, measure its current performance, analyse its failure points, and show that exceptions are rare and understood.
Practitioner guidance
- Define the workflow in measurable terms first. Rewrite each candidate automation into a precise problem statement, a start condition, a success measure, and a failure condition before any tool is configured.
- Measure the current process before redesigning it. Capture handoff delays, exception counts, manual touchpoints, and approval latency so you can prove whether automation improves the workflow or just accelerates noise.
- Analyse root causes, not symptoms. Separate process friction caused by missing data, fragmented ownership, or disconnected triggers from the visible delay that leadership wants fixed.
What's in the full article
JumpCloud's full blog post covers the operational detail this post intentionally leaves for the source:
- A step-by-step DMAIC application for workflow redesign before automation.
- Practical automation recipes for identity, device, and access operations.
- Scoring models for deciding which workflows are ready to automate.
- Financial modelling guidance for evaluating automation trade-offs.
👉 Read JumpCloud's analysis of why broken processes break automation →
Broken workflows and automation: what IT teams miss first?
Explore further
04/07/2026 10:06 pm
Automation exposes process debt before it delivers efficiency. The article’s core lesson is that workflow automation does not fix broken operational design, it makes the break visible at scale. That is why undocumented steps, weak handoffs, and unmeasured exceptions cause so many automation programmes to fail after deployment. For identity teams, the practitioner conclusion is simple: process maturity is the prerequisite, not the by-product, of automation.
A few things that frame the scale:
- 88.5% of organisations acknowledge that their non-human IAM practices lag behind or are merely on par with their human identity and access management efforts, according to The 2024 Non-Human Identity Security Report.
- Only 19.6% of security professionals express strong confidence in their organisation's ability to securely manage non-human workload identities, which shows how often control maturity is overstated.
A question worth separating out:
Q: What should teams do when automation starts creating more exceptions?
A: Treat the exceptions as evidence that the underlying process is unstable. Pause expansion, identify which data, handoff, or approval step is failing, and redesign the workflow before adding more automation. A rollback plan and audit trail should be in place before the next deployment.
👉 Read our full editorial: Automation magnifies broken processes, not operational discipline
