They ask AI where AI could help next instead of waiting for use cases to occur to them
Chapter Progress: Early DraftIn the fields of observation chance favors only the prepared mind.
Friction and imagination both point to use cases. Power users do not wait for obvious AI tasks. They expose a slice of their work and ask where friction, repetition, and delay suggest AI leverage, and they ask the playful question too: wouldn't it be cool if AI could do this thing I never thought to ask for. Since AI will over-suggest, you can have it triage its own list by frequency, value, judgment risk, data sensitivity, setup cost, and reversibility.
They give AI the raw context around the task
Power users open AI with a week of work in mind, not only a single task. Use-case discovery works best when the input is unpolished: a calendar, a task list, meeting notes, a backlog, a description of recurring frustrations, or a folder of repeated deliverables.
The core move is simple: 'Here is what I did this week. Identify where AI could have helped before, during, or after the work.' The model scans for patterns you are too close to see. The repeated email you rewrite every Monday. The meeting notes that never become action items. The claims in your research briefs that you always mean to verify. The internal tool you keep wishing existed.
This is applied to your whole workflow. Instead of asking AI to complete one task, you ask it to audit all your tasks for AI-eligible work. The discovery surface expands from 'this one thing I thought of' to 'everything I did this week.'
They turn repeated friction into experiments
Every annoying recurring task is a candidate for AI support. Power users feed those annoyances to AI and let it surface 'I hate doing this' as a clue. That clue points toward a possible experiment.
'I rewrite the same client explanation every week.' That is a prompt template. 'I lose time turning notes into action items.' That is a voice-to-output workflow. 'I keep forgetting which claims in my research briefs are verified.' That is a . 'I have ideas for internal tools but never turn them into specs.' That is a feature imagination session.
Friction is one of several feeds here, alongside curiosity and the things you wish existed. Power users keep a running list of tasks that annoy them, take too long, or feel mechanical, and they let AI keep that list current by scanning new work for the same patterns. Each entry is a candidate for one of the four tiers of AI opportunity. Every backlog item deserves a 15-minute safe-to-fail probe before it becomes a system: try the workflow manually once, observe where it creates value or risk, and only then encode it into a saved prompt, standard, workflow, or agent.
They ask for use cases in tiers
Asking AI for a random list of 'AI ideas' tends to produce shallow suggestions. Power users structure the request into tiers that map to increasing levels of investment:
- Quick wins: Useful today, no new tool needed. A better prompt, a different way to phrase the request, a context block that produces better output.
- Reusable prompt and standard opportunities: Repeatable task worth saving. A prompt template, a context block, a quality standard, or a skill that turns a one-time win into a recurring advantage.
- Workflow redesigns: Multi-step process that should be resequenced. AI changes which tasks should happen first, which can run in parallel, and which steps can be compressed or eliminated.
- Feature, agent, and build opportunities: Something that could become a tool, script, dashboard, automation, or agent workflow. More investment, more potential return.
This tiered structure connects to several later chapters. Quick wins connect to and focused dialogue. Prompt and standard opportunities connect to saving reusable prompts and standards. Workflow redesigns connect to redesigning workflows around AI. Feature and agent opportunities connect to building agents. Your use-case backlog gives you a personal short list of which of those chapters to read first.
They filter ideas by value, risk, and reversibility
AI will propose too much, so give it a triage grid and have it score each suggestion against six questions:
This keeps the chapter honest. High-frequency, low-risk, easily reversible tasks go first. High-stakes, high-sensitivity tasks go last, or never. The triage grid prevents the over-automation trap: automating everything AI can do regardless of whether AI should do it.
They imagine features before choosing tools
Power users ask a prior question before 'what app should I use?': 'What capability do I wish existed inside my work?' Because that question is tool-agnostic, it tends to surface the underlying need before the implementation choice narrows the options. This is the playful, imaginative feed at work: not every use case starts as a complaint, some start as 'wouldn't it be cool if.'
A button that turns meeting notes into a client-ready update. A dashboard that shows which projects are blocked. A that remembers your proposal style. A script that extracts every open decision from a folder of docs. An internal tool that turns a voice memo into a feature .
Power users use AI to imagine lightweight tools, internal features, automations, and agent workflows they could build or prototype, then climb a level and ask AI to draft the specification for the one worth building. The later chapters on building agents and on redesigning workflows around AI give them the means; this chapter gives them the habit of looking.
They test one small version before building a saved prompt or workflow
A power user does not build a saved prompt, workflow, or agent first. They run the idea once by hand in chat, then save the pattern if it works. Pick one suggested use case from the backlog, the smallest and most reversible one, and run a tiny version manually. If the voice-to-action-items idea looks promising, try it once with a real meeting recording. If the claim-verification idea looks useful, run it on one research brief.
One successful manual test is worth more than a detailed automation plan. It confirms the value before you invest in a saved prompt, standard, or agent. It also gives you the prompt, the context, and the quality standard that the later agent setup will need. From there you climb a level: capture that test into a prompt, standard, or workflow, then let AI help you evolve it as your work changes. The chapter on saving reusable prompts and standards shows you how.