The Sunday Work Reset One AI decision. One useful resource. One clearer workweek.

WHAT CHANGED?

You tried AI on a task, it worked once, and then it stalled. So you looked for the reason. Maybe it was the wrong tool. Maybe it was a bad prompt. Maybe you needed better training. You switched tools or bought a prompt pack, and the work did not get easier.

The reason is simpler than any of that. The task you handed over was already broken.

Here is what that looks like in practice. Say you want AI to write your customer proposals. You paste in a past proposal, describe the new job, and ask for a draft. What comes back is close, but the pricing logic is off, the scope section is missing the thing you always add verbally, and the tone is not yours. So you rewrite it. It takes forty minutes instead of the hour it used to take, and you tell yourself that is a win.

It may be a small improvement, but it is not yet a reliable workflow. You just did the whole job with an extra step in the middle.

The reason it went that way is that your proposal process was never written down. You have been making three judgment calls on every proposal without noticing you were making them. What you charge for travel. What you include for a repeat customer. What you leave out because this client always pushes back on it. Those calls live in your head. You never gave them to the AI, because you never said them out loud to yourself.

Write those three rules down, give them to the AI, and the draft has a real chance of coming back usable instead of becoming another round of rewriting.

If your quoting process lives in your head, AI cannot quote for you. If your onboarding steps change every time you run them, AI cannot write your onboarding email. If nobody has agreed what a good proposal looks like, AI will hand you back five versions and you will spend an hour picking one. That is not time saved. That is work moved.

AI did not fail at the task, it actually did the task you asked. The problem is the task itself.

What this changes:

Stop shopping for a better tool. Start looking for a task that is already clear enough to hand off.

Who should pay attention:

Anyone who has tried AI and watched it stall, and any manager whose team is using AI with no time saved to show for it.

What to do this week:

Pick one task you repeat. Write down the steps as they actually happen, not as they should happen. If you cannot write the steps, that is the task to fix, not the task to automate.

The Handoff Test

Five questions. Run any task through it before you give it to AI. Takes five minutes.

1. Can I explain this task to a new hire in under two minutes?

If the answer is no, AI will not get it either. Fix the task first.

2. Do I know what a good result looks like?

If you cannot describe good, you cannot check the output, and you will end up rewriting it yourself. That is not help.

3. Do I do this more than twice a month?

A one-off is not worth setting up. Repeat work is where the hours live.

4. What information does this task need, and can I hand it over safely?

Customer names, pricing, contracts, health details, employee records. If it is sensitive and you have no rule for it, set the rule before you set up the task.

5. Who checks the output before it goes out?

Name one person. If the answer is nobody, you do not have a workflow. You have a risk.

How to score it:

A task that passes all five is a strong candidate for AI support.

If it fails Question 1, 2, or 5, fix the process before you use AI.

If it fails Question 3, it may not repeat often enough to be worth setting up.

If it fails Question 4, stop and decide what information is safe to use before you put anything into an AI tool.

How to run it:

Take the task you wrote down Monday. Answer all five questions on paper. Do not skip question four because it feels like the boring one. It is the question that prevents a helpful shortcut from becoming an avoidable problem.

I'm hosting a free live ChatGPT lesson called:

"Stop Making ChatGPT Create More Work."

It will help you identify one task ChatGPT can actually help with, understand why it fits, and set one boundary before you start.

Registration opens this week. Reply WORK to this email and I'll send you the link before I post it publicly.

How I help:

AI Advisory. Executive alignment, AI roadmaps, and decision frameworks for leaders navigating AI.

AI Governance. Safe-use policies, risk review, and governance playbooks built for organizations where the cost of getting it wrong is real.

AI Training & Adoption. Role-based training and playbooks that hold after the rollout ends.

AI Speaking. Signature keynotes (The Operator's Edge, The Hidden Hour), executive briefings, and workshops.

Reach me: keenallianceconsulting.com set a time or reply to this email.

See you next week!

Andrew Keener 

Founder, Keen Alliance Consulting 

Follow Andrew Keener for where AI belongs, and where it doesn't.