Evidence-Based Decisions Reduce Assumption Risk
- Decisions sit further and further from current conditions.
- Prediction has replaced observation, and errors scale faster than learning catches up.
Intent Drift
Unintended System Conditions. A USC is an unplanned state that builds up when principles are repeatedly violated. It is the default output of the system as currently designed, not a one-off incident or a people problem. Fixes aimed at symptoms improve things briefly before the condition reasserts itself. Which principles feed each condition is NOT stored here: it is derived at runtime from the consequence-to-condition links in the principle files.
Intent Drift
The system keeps executing against goals or specifications that no longer reflect reality. Outputs stay internally consistent but drift from actual needs.
The team built exactly what was asked. It is no longer what is needed.
Read more
The delivery system continues executing against goals, priorities, or specifications that no longer reflect current business reality. The original intent was reasonable when it was set. Over time, context shifted, understanding deepened, or business needs evolved, but the governing intent was never updated. Teams produce outputs that are internally consistent with what was asked but progressively misaligned with what is actually needed. The condition develops gradually and without obvious signals because delivery continues normally. The gap between intent and reality only becomes visible when outcomes fail to produce the expected value.
In agentic delivery
In human-driven delivery, Intent Drift surfaces gradually through human friction. A product manager notices the roadmap feels wrong. A developer questions whether the feature is still needed. Those friction signals create natural correction opportunities even when no formal process exists to act on them. In an agentic system those signals disappear entirely. Agents execute continuously and confidently against whatever intent they were given. There is no frustration, no hesitation, no quiet doubt that something feels off. The drift is silent until the outcome is wrong.
- Agent pipelines continue executing against goals that leadership updated weeks ago because nobody translated the strategic change into updated agent instructions
- The volume and velocity of agent output creates an appearance of progress that masks the growing gap between what is being built and what is actually needed
- Intent specifications governing agent behavior have not been reviewed since initial deployment despite significant changes in business context
- When outputs miss the mark, the instinct is to improve the agent rather than revisit the intent the agent was given
Principles that feed this condition
Primary
Progressive Commitment Reduces Uncertainty Exposure
- Sunk cost and fear of looking wrong keep money flowing to weak bets.
- Output keeps rising while outcomes drift.
- Wrong assumptions and misaligned agent output scale at machine speed before anyone notices.
Continuous System Learning Sustains Performance
- Outdated practices stay embedded, and AI now repeats them faster.
Clear Operating Rules Grounded in Intent Enable Autonomous Decisions
- Rework comes from the wrong goal done well, yet gets treated as execution failure.
- Agents repeat mechanical compliance and misread intent at scale.
Automate by Task Nature, Not Capability
- Verification is bolted on afterwards to tasks never designed to be checked.
- Judgment calls move into AI workflows where nobody is set up to question output.
Agent Trust Must Be Continuously Earned, Not Historically Assumed
- Agent outputs pass every check but no longer fit what the organization needs.
Secondary contributors
Making System Conditions Visible Enables Timely, Informed Decisions
- Leaders make decisions with a confidence the facts don't support.