RIGHT ENGINES
A critical project is late and nobody agrees why
A practical guide to diagnosing a critical project is late and nobody agrees why and deciding whether an AI engine is justified.
The problem
The visible delay is often a symptom of unclear ownership, blocked dependencies, or an unrealistic critical path.
Symptoms
Look for recurring delay, correction, escalation, or paid attention around this workflow. Confirm the pattern with current plan, milestones, dependencies, owners, decisions, risks, and available capacity. rather than relying on anecdotes.
Why it happens
Status updates report activity rather than the constraint controlling delivery.
What it costs
Measure the burden created by a critical project is late and nobody agrees why: responsible-person hours, elapsed delay, rework, error exposure, outside spend, and decisions deferred. Use a representative baseline; do not invent ROI.
What people usually try
More meetings, longer status decks, or arbitrary deadline resets.
Where AI helps
For this problem, AI may help by turning current plan, milestones, dependencies, owners, decisions, risks, and available capacity. into a constraint diagnosis, evidence-backed critical path, recovery options, owners, and dated checkpoints. It should preserve evidence and expose exceptions for review.
Where AI does not help
Do not automate political accountability or conceal uncertainty behind a confident forecast.
What an engine could do
A constraint diagnosis, evidence-backed critical path, recovery options, owners, and dated checkpoints.
Inputs required
Current plan, milestones, dependencies, owners, decisions, risks, and available capacity.
Success measures
Accepted recovery plan, resolved blockers, milestone reliability, and decision turnaround.
Risks
The specific failure boundary is clear: Do not automate political accountability or conceal uncertainty behind a confident forecast. Also test provenance, access, false confidence, and the effect of unreliable inputs.
Small / medium / large solution
Start with the smallest validated route. Use an Outcome Kit only where its stated inputs and tests fit a critical project is late and nobody agrees why; adapt a proven engine where workflow context differs; commission private work only when the evidence, integration, governance, or rights justify it.
Estimate this problem’s burden · Decide whether it is worth building for · Review reusable Outcome Kits
Published 2026-08-31 · Updated 2026-08-31 · Version 1.0