Intent Drift · Secondary
The team built exactly what was asked. It is no longer what is needed.
How this principle feeds it
confident leaders keep commitments running against assumptions that no longer match reality.
Principles Library / Visible Conditions
Making System Conditions Visible Enables Timely, Informed Decisions
Make the conditions shaping system performance visible, and build the ability to trace decisions to their causes before a system operates, not after a failure.
Decisions grounded in shared, traceable evidence produce better outcomes than decisions grounded in assumptions, filtered reporting, or records reconstructed from memory after a failure. Visibility should extend beyond progress and status to include workload, dependencies, risks, uncertainty, constraints, decision assumptions, quality signals, and emerging outcomes.
Risks and bad news reach the people who can act while there are still options.
Leaders decide from shared, traceable facts instead of filtered reports or assumptions.
When something fails, you can trace it to the decision that caused it and fix that.
The team built exactly what was asked. It is no longer what is needed.
confident leaders keep commitments running against assumptions that no longer match reality.
Customers find defects before the team does.
problems found late are found after they are cheap to fix.
Every quarter starts over. Nothing compounds.
each late crisis forces a replan, so commitments are dropped before they finish.
Every team hits its targets. The product still misses.
without a trusted shared view, each group optimizes the numbers it can see, and end-to-end value fragments.
We know what happened. Nobody can explain why.
outcomes are recorded but cannot be traced to the decisions that produced them.
The review process still runs. It stopped meaning anything.
review continues while the volume and opacity of agent decisions outrun it.
The agents are building. Nobody is sure it still matches the plan.
untraced agent substitutions become the new baseline without anyone flagging the divergence.
AI signs are on by default. Use the AI switch to exclude them.
For each consequence: Not seen / Sometimes / Persistently / Not sure.
Based on your answers, recognize at most two conditions: No / Partly / Yes, that is us.
Intent Drift: The team built exactly what was asked. It is no longer what is needed.
Quality Fragility: Customers find defects before the team does.
Strategic Volatility: Every quarter starts over. Nothing compounds.
Local Optimization Bias: Every team hits its targets. The product still misses.
Attribution Failure: We know what happened. Nobody can explain why.
Oversight Erosion: The review process still runs. It stopped meaning anything.
Implementation Drift: The agents are building. Nobody is sure it still matches the plan.
What else have you observed? Optional. Shared only if you choose to start coaching.
Start a Diagnosis session: review and edit a draft containing your note, conditions you recognized in Step 3 (including whether they match partly), warning signs and consequence ratings before sending it. No scores or principle files are sent.
Showing all 7 practices. Answer the Diagnose questions and the most relevant come first.
Make queues, dependencies, risks and uncertainty visible, and get bad news surfacing early.
Start by asking your team: What do we know now that leadership can't see, and where does reported status differ from what people experience?
Risks known on the ground reach the people who can act, and teams raise problems before they escalate.
Design reports to show real system conditions, and keep what people disclose apart from how they're evaluated.
Start by asking your team: What incentives discourage people from surfacing bad news?
Status reports match what people actually see, and outcomes and reports move together.
Before a system runs, decide what must be recorded, at what level of detail, and for how long.
Start by asking your team: What would a regulator be unable to answer if they reviewed our current decision records?
Reviews can name the decision behind a failure, and a failed agent pipeline can be replayed step by step.
Design decision records that capture the inputs, the governing policy and the reasoning along with the output.
Start by asking your team: Who could reconstruct the reasoning behind the last few significant decisions this system made?
Post-incident reviews name the decision that caused a failure, and accountability follows who decided.
Label which signals are facts and which are interpretation or assumption, and make key inputs and trade-offs traceable.
Groups work from one shared view of the system, and reports mark which claims are assumptions.
In AI-augmented systems, make automated decisions, their inputs, actions, exceptions and outcomes observable and auditable.
Start by asking your team: As execution speeds up, which signals and decision paths must never be filtered, delayed or left unrecorded?
Leaders can see why an agent produced an output, and agent assumptions are open to inspection.
Before deployment, simulate an incident and confirm the causal chain can be rebuilt without relying on anyone's memory.
Start by asking your team: If something went wrong today, could you trace the decision that caused it to a specific point, input and governing policy?
After a simulated incident, the team traces the cause to a specific decision, input and policy.