🚩Tactical Memo 104: Faster Tasks Don’t Move the Deadline
AI saved you hours. The launch still slipped 3 weeks.
Justin Bateh
October 18, 2026
Tactical Memo #104 | Modern Work · AI | 4 min read
Tactical Memo: a new issue every Tuesday, Thursday, and Sunday on career growth, AI at work, and modern management, read by 18,000+ leaders at Amazon, Deloitte, and Google. Forwarded? Subscribe free.
Last day to save 20% Become an AI-Powered Project Manager. In 4 weeks, you'll build your AI Project Operating System, with agents handling project tasks, automation running workflows, and agentic AI helping manage the work with you. You'll leave with an AI Project Brain and 6 working AI agents built around real project work. Use code AIPM20 at checkout to save 20%. The discount ends tonight. |
In brief The situation: AI cut your task time, and your projects still run late. The mistake: Most people point AI at tasks, and the delay was never in the tasks. The move: Point AI at the whole project and the gaps between tasks. |
The story
Picture your last project.
AI wrote the status report in 2 minutes. It turned a 40-email vendor thread into 5 bullets.
You got hours back every week.
The launch still slipped 3 weeks.
The legal review sat in someone's inbox for 9 days. The budget call waited on a VP who was traveling. Design and engineering each thought the other team owned the spec.
None of that was task work. Your AI never touched any of it.
If AI saves you hours and your projects still finish late, you're in the same spot.
Using AI makes your tasks faster. Running the project with AI is what moves the deadline. Here's how to work with it.
The moves
1/ Find where the project waits
Pull up your last project that ran late.
Mark every stretch where nobody was working on it. Add up the waiting days and compare them with the working days.
The waiting usually wins.
That's the part of the project task-level AI never sees.
Find the waits first, because that's where the weeks go.
Checklist Which decision did we wait on longest? Which approval sat the longest, and with whom? Which handoff did 2 teams each think the other owned? |
2/ Give AI the whole project
Build one running file for the project. Goals, owners, dates, decisions made, and decisions still open.
Feed it every update and every set of meeting notes. Then ask your questions about the project as a whole.
A tool that sees 1 email can only improve the email. A tool that sees the whole project can tell you what's about to slip.
That running file is the difference between using AI and running the project with it.
Ask this "Based on everything in this project file, what is most likely to slip in the next 2 weeks, and which open decision is it waiting on?" |
3/ Log every decision the day it gets made
When a call gets made, it goes into the project file that day.
Who decided, what they decided, and what work it changes.
Teams lose days arguing again about calls nobody wrote down.
A written decision is one the team stops reopening.
Do this After each meeting, paste the notes in and have AI list every decision, who made it, and what work it changes. You confirm each one before it goes in the file. |
4/ Send the follow-up within the hour
Have AI draft the follow-up the minute a meeting ends, with an owner and a date on EVERY action.
Check it, then send it before you start your next task.
People forget what they agreed to once they're back in their inbox.
Send the follow-up while everyone still remembers saying yes.
Say this "Here's what we agreed, who owns each item, and the date it's due. Reply today if anything is wrong." |
5/ Flag open decisions before they get old
Have AI flag any decision that's been open more than 3 days.
Each one goes back to its owner with a date attached.
Nobody notices a decision sliding until the work behind it stops.
An open decision with no date is how 3 days turns into 3 weeks.
Do this Every open decision gets 3 things in the project file. The owner, the date it's due, and the date it was first raised. |
6/ Get a waiting list every morning
Have AI send you a short list each morning of who is waiting on whom.
Clear the top item before you open anything else.
Gaps close when someone chases them. Most project managers run out of time to chase.
AI doesn't.
Hand AI the chasing, and spend your time on the calls only you can make.
7/ Ask for the risk you don't want to hear
Once a week, ask the project file what could make you miss the deadline.
Push past the first answer. Ask what you're assuming that might be wrong.
Status reports lean green because nobody wants to send bad news. A tool reading the whole file gives you the same answer either way.
The earliest you hear about a risk is the cheapest time to fix it.
Ask this "What in this project file suggests we'll miss the deadline, and what am I assuming that could be wrong?" |
The outcome
| Most people | You |
|---|---|
| Use AI to write the update faster | Use AI to spot what's about to slip |
| Save hours on tasks | Save weeks on the schedule |
| Chase open decisions from memory | Track every open decision with an owner and a date |
Proof. Atlassian surveyed 12,000 knowledge workers. They said AI makes them 33% more productive, yet 96% of companies had not seen dramatic gains in company-wide efficiency. Atlassian, 2025
This week
Your project was NEVER short on fast tasks.
The people who finish on time use AI to track the whole project and chase the gaps between the work.
Try it next time a project starts to drift. Ask your AI what the project is waiting on.
A project can finish late with every task done early.
Reply
Where does your current project wait the longest? Reply with decisions, approvals, or handoffs. I read every one.
Until next time.
Justin Bateh, PhD
CEO @ AI Operators Lab
Founder and Editor, Tactical Memo
LinkedIn Top Voice | 8x Awards