A conversation with the Action Plan team
How do I get started with AI agents?
Everyone says "use AI agents." Nobody says where to actually start. I don't want to spend a month configuring something just to find out it doesn't fit how I work.
Start with one problem that's painful, repeatable, and checkable — not a platform rollout. State the goal in plain words, let the system propose a plan with a cost, approve it, review the result. If your first win takes more than a day, the tool is wrong or the problem is too big. Onboarding should feel like a conversation, not a configuration.
The "painful, repeatable, checkable" test does most of the work, so let me unpack it:
- Painful — you feel it weekly. Email triage, follow-ups you forget, the report you dread, the app you keep meaning to have built. Pain keeps you engaged through the calibration week.
- Repeatable — it happens again and again, so every improvement compounds. One-off tasks teach the system nothing.
- Checkable — you can verify the result in minutes. That's how trust gets built: small verified wins, not promises.
What not to start with: your most complex process, anything irreversible, or "automate everything." Those come later, on top of earned trust.
And here's a quiet rule for judging any AI tool: if it needs prompt engineering from you, it's pushing its work onto you. You should describe outcomes in plain language. Turning that into plans, clarifying questions, and verified work is the system's job. In Action Plan, your first hour looks like this:
- 1 Tell me the goal, in your words. I'll ask the two or three questions that actually matter.
- 2 I bring you a plan: steps, which specialist does what, and what it will cost. You adjust and approve.
- 3 The team executes behind quality gates. You review a real, verified result — today, not "after setup."
One more honest note: the first week has a learning curve in both directions — you're learning what to delegate, and the team is learning your standards. The systems that win aren't the ones with zero friction on day one; they're the ones where every correction you make is remembered, on the record, so you never make it twice.
If you're not sure which problem to start with — that's literally what the message box below is for. Tell me what's eating your week, and that becomes your first conversation when your invite opens.
Getting started with AI agents
- Pick one problem that's painful, repeatable, and checkable.
- State the goal in plain language — a good system asks the clarifying questions itself.
- Approve a plan with a visible cost — never a black box.
- Review a verified result the same day — trust is built on small checked wins.
- Expand only on earned trust — complex and irreversible work comes later.
Good first projects
Email triage, quote follow-ups, a weekly report, a simple internal app — anything you can verify in minutes and feel within a week.
How long should setup take?
First win in a day. If it takes a month of configuration, the tool is doing onboarding wrong.
Action Plan is in private preview. When your invite opens, onboarding starts from exactly this conversation.
Got it — you're on the early-access list, and I've saved your goal: . When your invite opens, this is the first thing we'll work on together. Watch your inbox.