Course 3 - Module 5 - Lesson 1
Spot new AI opportunities by watching where friction repeats
Your AI Opportunity Map
Notice friction before proposing a solution
Friction is the repeated effort that makes work slower or more confusing. It might appear as copying details between systems, searching for the same policy, rewriting unclear notes, or waiting for a missing field. Keep a short diary of specific incidents instead of a general wish for more automation.
For each incident, record the task, frequency, minutes lost, cause, and people affected. 'Spent twelve minutes locating the approved room layout again' is useful evidence. 'Need an AI platform' skips the problem and makes it harder to compare simpler fixes.
Look for the real source of effort
Three repeated searches for a layout may mean the file is badly named, not that you need an agent. A shared link and clear owner might solve the problem. AI is more promising when varied material needs interpretation, summarization, or drafting that remains easy to review.
Group diary entries by cause, such as missing information, repeated transformation, or unclear responsibility. Ask a colleague whether the same issue affects them. More people experiencing a problem may increase its value, but it may also reveal permissions and ownership issues you need to resolve.
Turn one observation into a testable opportunity
Write an opportunity statement: 'If we produce a checked room brief from confirmed booking details, coordinators may spend less time rewriting notices.' State a baseline, expected improvement, and failure condition. For example, the trial fails if any draft changes a confirmed room or time.
Choose one small experiment using sample material and measure preparation, review, and corrections. Keep the option to stop. Evidence that a simpler template solves the problem is a useful result, even when it means no AI integration is needed.
Tiny experiment
Record three imaginary friction incidents, group them by cause, and turn one into an opportunity statement with a baseline and a clear failure condition.
Confidence check
What observation would convince you that the best fix is a clearer process rather than a new AI tool?