Step 1: Analyze Workflows for AI Opportunity

Part of: AI Workflow Framework

What This Is

A structured audit that helps you find where AI fits in your work. The analysis supports two lenses — perspectives you can analyze from: Individual (workflows you personally perform) and Organizational (value chain processes that deliver on business objectives). The AI scans what it already knows about you, asks which lens to use, interviews you with lens-specific questions, then analyzes the results to identify opportunities you’d miss on your own.


What you’ll do	Walk through a guided conversation covering your role, tasks, and pain points
What you’ll get	Three to five candidate workflows registered in your registry backlog, a prioritized report, and one recommendation to build first
Time	~15–20 minutes

Why This Matters

Most people adopt AI by reacting to problems — they reach for ChatGPT when they’re stuck on an email or ask Claude to summarize a document. That’s useful, but it misses the bigger picture.

A proactive audit of your workflows can reveal opportunities you’d never notice in the moment: repetitive tasks that could run on autopilot, decisions that would benefit from an AI collaborator, and multi-step processes that could be orchestrated end-to-end.

This step guides an AI through a structured analysis of your work — from either an individual perspective (your personal tasks and pain points) or an organizational perspective (your business’s value chain and strategic processes) — and produces a prioritized report of opportunities. Each one names the work, the pain, the kind of value AI adds (Streamline, Automate, Accelerate, or Create value), what gets produced, and how much it matters.

Analyze stays strategic. It names the work, the pain, the kind of value AI adds, and what gets produced — never the tool, connector, or integration that would do it. Those are decided in Design (Step 3). If a report starts naming products or connectors, say: “Stay strategic — Design picks the tools,” and the skill will restate the recommendation as an outcome.

How the Skill Works

The skill runs six phases in order:

Read what’s known — The registry (your business, processes, existing workflows) plus memory and project files. Presented for you to confirm; nothing you’ve already said is asked again.
Lens selection — Individual (your own workflows) or Organizational (your business’s processes). Inferred and confirmed when obvious.
Discovery interview — With a registry, a walk through your processes asking where the friction is, tied to your objectives. Without one, focused questions, one at a time, only on what phase 1 didn’t cover.
Opportunity report — A summary table and detailed cards, grouped by value lever, ordered by impact.
Candidates registered — You pick up to five; each becomes a backlog Workflow node in your registry, filed under one of your processes, and the skill recommends which to build first (a starter-sized one).
The other lens — Offered as a later session.

Most people discover 5–15 opportunities across the four value levers. Pick up to five to start with.

How to Use This

This step is facilitated by the analyze AI Workflow Framework skill. How you get it depends on your platform — see Install the Hands-on AI Plugin for installation instructions.

How to start: Say “run the analyze skill” (or “analyze the workflow”) — works on every platform. With the plugin installed, Claude Code also accepts /handsonai:analyze, and Cowork lists it when you type /.

Platform compatibility: Claude (Chat, Cowork, Code) ✓ | ChatGPT & Codex ✓ | Gemini (Spark, Enterprise, CLI) ✓ | M365 Copilot ✓ | Cursor / Antigravity ✓

Start with this prompt:

I'd like to analyze my workflows for AI opportunities. Help me audit
what I do and identify where AI could help.

The skill runs the six phases above and produces a prioritized opportunity report, and registers your chosen candidates in the registry backlog.

Example prompts

"Help me analyze AI workflow opportunities"
→ Runs the full audit and produces a prioritized opportunity
  report with structured workflow candidates

"I want to figure out which parts of my job could benefit from AI"
→ Interactive discovery session followed by a structured report
  with specific, actionable recommendations

How to Prioritize

Start small if you’re new to AI — a workflow of roughly 3–5 steps, touching one system, that you start by hand. It teaches the full loop with the least at stake.
Follow the value lever you’d use to justify it — an Automate opportunity pays back in hours you stop spending; an Accelerate one in cycle time; a Create value one in something that didn’t exist before.
Come back for the ambitious ones — opportunities where the AI would need to decide more of what happens next are worth building once you’ve shipped a small one. Design (Step 3) is where that question gets settled.

What Analyze Produces

The AI Opportunity Report (ai-opportunity-report.md) captures:

Report header — your name, role, date, opportunity count, and top recommendation
Summary table — every opportunity listed with its value lever and impact level
Top recommendations — the 3 highest-priority opportunities with one-sentence rationales
Detailed opportunity cards — grouped by value lever (Streamline → Automate → Accelerate → Create value), each with: why it’s a good candidate, current pain point, how AI helps, the value lever, and what changes for the business
Workflow Candidate Summary — structured metadata for the workflows you choose to pursue: name, description, trigger, deliverable, value lever, pain point, AI opportunity, frequency, priority, reasoning, and lens. Organizational-lens candidates also include business objective, stakeholders, and success metrics.
Backlog Workflow nodes — one per chosen candidate in registry/workflows/, linked to the report. Deconstruct picks one up from here.

The Workflow Candidate Summary is the input for Deconstruct Workflows (Step 2) — the trigger and deliverable fields map directly to the scope check that starts the deconstruction. Its business objective and success metrics also seed Deconstruct’s Value & Measurement section, so an organizational-lens candidate arrives with two of those answers already drafted.

Analyze does not settle the question of value, and shouldn’t. It surveys many workflows to find the few worth pursuing; the outcome that changes, and the number that measures it today, are established once a specific workflow is on the table. Deconstruct asks for both.

See three complete example reports to get a feel for the format and level of detail.

Tips for Better Results

Bring your registry. Analyze is sharpest when it can walk your real processes. Set it up first with AI Registry Setup — about 25 minutes, once — and run Analyze straight after; the setup interview offers to.
Use a tool with memory or projects enabled. The richer the AI’s context about your actual work, the more specific and useful the recommendations will be.
Be concrete when answering questions. “I spend 30 minutes every Monday formatting a status report from three Jira boards” is far more useful than “I do reporting.”
Run it again in a few weeks. As you have more conversations and the AI learns more about your work, re-running this prompt will surface new opportunities.
Share the output with your team. Some of the best opportunities come from workflows that span multiple people — your colleagues may see possibilities you don’t. The organizational lens is especially useful for this.
Try both lenses. The individual lens surfaces your personal pain points; the organizational lens surfaces value chain opportunities that may have higher strategic impact — worth a separate short session.

Example Reports — three complete AI Opportunity Reports (Marketing Ops Manager, AI Instructor, and VP of Operations using the organizational lens) showing what the finished deliverable looks like
Next step: Ready to act on an opportunity? Use the Deconstruct Workflows guide (Step 2) to break it into discrete steps.
AI Workflow Framework — the full seven-step methodology
Prompts
Agents