Eight questions. A specific answer about which element of Intent Management™ is failing for your team right now — and what to do about it.
Question 1 of 8
Question 1 of 8
When your team finishes a piece of AI-assisted work, how often does the output miss what you actually needed?
Question 2 of 8
If you asked two people on your team to describe what "done" looks like for a current project, would they give you the same answer?
Question 3 of 8
When your team reviews AI-generated output, what does the review process actually look like?
Question 4 of 8
When an AI output gets rejected or significantly reworked, how clearly can your team explain why?
Question 5 of 8
When AI-generated work needs a significant correction, what typically happens?
Question 6 of 8
How would you describe your team's relationship with the decisions you've delegated to them?
Question 7 of 8
How does your team know what's off-limits when working with AI tools?
Question 8 of 8
When priorities or conditions change, how quickly does your team's AI-enabled work adjust?
Your Primary Gap
Outcome Definition
Your team is working hard toward a target that hasn't been made explicit.
The most common version of this: everyone understands the general direction, but no one has written down what "done" actually looks like — specifically enough to test. AI tools amplify this problem because they produce output faster than humans can notice the drift.
The fix isn't more communication. It's more precision. Before the next project starts, write down what success looks like in terms that your team could use to evaluate their own work — without checking with you.
Where to Focus First
Before any AI-assisted work begins, write a one-paragraph outcome statement — specific enough that two people would agree on whether the output meets it
Test it by asking a team member: "What would success look like?" If they can't articulate it back, the statement isn't clear enough yet
Review Chapter 3 of Alignment at Speed for the Outcome Definition framework
Try the Week 1 experiment in the First 90 Days guide — one decision, fully written out, before work starts
Outcome Definition is the foundation everything else builds on. Getting it right on one project this week is more valuable than reading about it for another month.
Your team knows roughly what good looks like — but can't apply it consistently.
This shows up as: different reviewers applying different standards, long debates about whether output is "good enough," and AI tools getting blamed for quality problems that are really calibration problems. The output isn't wrong — the evaluation framework is missing.
The fix is making your quality standards explicit before output is produced, not after. When the criteria exist only in people's heads, consistency is impossible — and AI tools have no way to orient toward what you actually want.
Where to Focus First
For one current project, write down 3–5 specific criteria that separate acceptable output from unacceptable — concrete enough that two reviewers would reach the same verdict
When options compete, name the trade-off explicitly: speed vs. quality, completeness vs. simplicity — and decide in advance which governs
Review Chapter 4 of Alignment at Speed for the Evaluation Criteria framework
Share the criteria with your team before review, not during it
Most AI quality problems aren't AI problems. They're calibration problems. Getting your evaluation criteria explicit is the fastest path to consistent output.
Your team doesn't know who actually gets to say yes — so everything comes back to you.
Decision Authority is the most commonly overlooked element — and the one that creates the most friction when it's missing. When authority isn't explicit, teams default to escalation. Escalation creates bottlenecks. Bottlenecks slow AI adoption to the pace of one person's capacity.
The fix isn't trusting your team more. It's being explicit about what decisions they're authorized to make — and then honoring that authorization even when you would have decided differently.
Where to Focus First
For one current project, write down: what decisions can the team make without me, what requires my input, and what requires escalation beyond me
When you give someone decision authority, mean it — stepping in anyway teaches them they still need your approval
Review Chapter 5 of Alignment at Speed for the Decision Authority framework
Check whether your team believes they're allowed to use the authority you've given them — sometimes explicit permission needs to be restated
If you're the bottleneck for most decisions in an AI-enabled workflow, the constraint isn't the AI. It's the authority structure. That's fixable in a single conversation.
Your team is operating without a clear map of what's off-limits.
Constraint Boundaries are the non-negotiable limits — regulatory requirements, budget ceilings, authority structures, organizational policies — that govern what AI-enabled work can and cannot do. When they're not explicit, teams discover them the hard way: by crossing them.
This is especially costly with AI tools, which can produce output that violates constraints efficiently and at scale. The fix is mapping the limits before work starts, not after the violation.
Where to Focus First
For one current project, name 3–5 things that would make the output unacceptable regardless of quality — regulatory limits, cost ceilings, stakeholder sensitivities
Write constraints down as explicit rules, not informal understandings — "we don't" is not a constraint; "outputs may not include X" is
Review Chapter 6 of Alignment at Speed for the Constraint Boundaries framework
Share constraints with the team and with any AI tools or prompts being used — if the constraint isn't in the system, the system will ignore it
Constraint Boundaries are the limits that can't be traded away. Getting them explicit before work starts is the difference between a preventable violation and one you're managing after the fact.
Your intent was clear when you set it. It's drifted since then.
Communication Cadence is what keeps the other three elements calibrated over time. Intent drifts — priorities shift, conditions change, and AI-enabled work continues in the old direction until someone notices. The gap between when intent was set and when it gets refreshed is where most accumulated rework lives.
The fix isn't more meetings. It's structured checkpoints — deliberate moments where the team asks whether the original outcome, criteria, and constraints still apply given what's changed.
Where to Focus First
Build a lightweight intent review into your existing rhythm — weekly or at project milestones — specifically asking: does our original outcome still apply?
When conditions change significantly, treat it as a trigger to explicitly revisit intent rather than assuming the team will adjust on their own
Review Chapter 7 of Alignment at Speed for the Communication Cadence framework
Read the Beyond 90 Days guide for how to sustain the discipline as conditions evolve
Most teams set intent once and assume it holds. It doesn't — especially with AI tools moving fast. Building the cadence to refresh it is what separates a one-time practice from a lasting discipline.