AI Adoption Filter

D06 - Technology Coherence

Definition

A four-question test applied before automating any operational task with AI tools, evaluating whether the task eliminates enough time, requires intelligence rather than judgment, produces usable output without heavy editing, and fails in a predictable way.

• Does the task consume more than two hours per week for the team?
• Is the task intelligence work — data retrieval, organization, and summarization, rather than judgment work?
• Is the output usable with less than 20% editing before it reaches a leadership document, a vendor, or a team member?
• Are the tool's failure modes predictable enough to be caught in a systematic review?

Most AI automation failures happen because the wrong task was automated, not because the tool was wrong. The filter exists to separate tasks that are genuinely safe to hand to AI from tasks that will fail silently and create more work than they save.

Indicators

• You've automated something that still requires significant editing before the output is usable.
• An AI-generated output has been shared or acted on before anyone verified its accuracy.
• Your team asks, "Should we use AI for this?" without any defined criteria for answering that question.

Quick Win

• Pick one task you are considering automating.
• Answer these four questions, yes or no:

1. Does it eliminate 2+ hours/week?
2. Is it intelligence-based, not judgment-based?
3. Is the output usable with <20% editing?
4. Does it fail predictably?

• If all four answers are yes, proceed. If any answer is no, do not proceed.

Assess

• Take the HQ Score Assessment

Apply

• The Headroom HQ AI Starter Kit

Related

• No-Code Automation
• Tool-Audience Fit
• Pilot-and-Report
• Tech Stack Tier
• The Four-Question Automation Filter