Read time: 10 minutes

Welcome to Tactical Memo, my newsletter where I share clear lessons and simple systems for project managers, operators, and team leaders navigating the AI era. Each issue tackles one real workplace challenge and hands you a ready-to-use solution.

If you want practical guidance you can use at work this week, you are in the right place.

👉 Why Read This Edition: You will learn why AI is making leaders dangerously comfortable at the 80% mark, and how the final 20% of any project, decision, or deliverable is where your reputation is actually built or broken.

FREE LIGHTNING LESSONS

I have two free sessions dropping this week. Each one is tight, tactical, and built to make you sharper.

Words That Work: Communication for Project Leaders

Friday, March 27 | 11:00 AM EDT | 30 minutes | Free

Hosted by Justin Bateh & Jill Avey

Most project leadership breakdowns are not due to poor strategy. They are due to poor conversation. In this session, you will learn how to use Conversational Intelligence to lead with influence, build psychological safety, and turn difficult conversations into opportunities for alignment. Register free.

How PMs Stay Irreplaceable as AI Changes Project Management

Friday, March 27 | 11:45 AM EDT | 30 minutes | Free

Hosted by Justin Bateh, PhD

AI is rewriting how projects are planned, executed, and led. Most PMs are still operating like it is 2018. This session gives you a clear, battle-ready game plan to stay indispensable in 2026 and beyond. Register free.

Seats are free. Excuses are not. Grab yours.

THE PROBLEM

I need to tell you about something that happened to me a few months ago. Because it changed how I think about AI and leadership. And I think it will change how you think about it too.

I was working with a leader who had just rolled out AI across his entire team. Smart guy. Ambitious. He was using AI for everything. Project plans, stakeholder updates, risk assessments, even the weekly team email. His output had tripled. His calendar had never been cleaner. On paper, he was the most productive leader in the organization.

Then I sat in on one of his project reviews.

A team member flagged a risk. A real one. The kind that does not show up in a dashboard. She said the vendor relationship was deteriorating and she was worried about a missed delivery in Q3. You know what the leader did? He pulled up his AI-generated risk register, pointed to a line that said vendor risk is low based on historical data, and moved on.

She stopped talking after that. Not just in the meeting. For the rest of the project.

That moment stuck with me. Because the AI was not wrong. Based on the data it had, the risk was low. But the person in the room had context the tool did not. She had been on calls with the vendor. She had heard the tone shift. She knew something was off. And the leader chose the polished output over the person sitting right in front of him.

That is the Last Mile Problem.

AI gets you 80% of the way there on almost anything. The draft, the analysis, the framework, the deck. That 80% is fast. It is clean. It is impressive. But the final 20% is where the real thinking lives. It is where context matters. Where judgment matters. Where courage matters. And I keep watching leaders skip it because the 80% looks so good they convince themselves the job is done.

It is not done. It is not even close. And I am telling you this because I have made this mistake myself. I have sent things out that were 80% AI and 20% nothing. And every single time, it cost me. Maybe not right away. But eventually, someone noticed. The work was clean but it was hollow. It did not have my thinking in it. It did not have me in it.

THE SOLUTION: THE LAST MILE FRAMEWORK

After that project review, I started paying attention. I looked at every deliverable I touched and asked myself where is the 20% in this? Where did I actually add something the tool could not?

That question turned into a framework I now use with every leader I coach. I call it the Last Mile Framework. Five things you do after the AI gives you the 80%. These are not optional. These are the difference between a leader who uses AI well and a leader who is being used by it.

1. Pressure-test the assumptions.

I learned this one the hard way. Early in my career, I inherited a project plan from a predecessor. It looked thorough. Timeline, milestones, resource allocation, all of it. I trusted it because it looked professional. Three weeks in, the whole thing collapsed. The resource estimates assumed full availability, but two of the key people were already committed to another initiative. Nobody had checked.

AI does the same thing your predecessor did. It builds on whatever you feed it and it does not push back. It does not say, hey, are you sure about that timeline? It does not flag that your cost estimate assumes a vendor rate from 2024 when the contract was renegotiated six months ago.

Your job in the last mile is to be the person who checks. Pull the assumptions apart. Ask what are we taking for granted here? Ask what would my toughest stakeholder challenge in this plan? I promise you, the five minutes you spend doing this will save you five weeks of cleanup later.

2. Add the context only you have.

Let me go back to that project review I mentioned. The leader had a risk register. It was thorough. It covered budget risk, timeline risk, technical risk. But it did not cover the one risk that actually mattered, which was the vendor relationship.

Why? Because AI does not know that your VP of Engineering is two weeks from burnout. It does not know that the client smiled during the demo but their follow-up email was unusually short. It does not know that procurement delayed the last initiative by six weeks and they are about to do it again.

You know those things. That is institutional memory. That is relationship intelligence. That is pattern recognition that comes from showing up every day and paying attention. The last mile is layering that knowledge onto the AI output. Without it, you have a plan that looks right on a screen but falls apart in the real world.

3. Make the call nobody wants to make.

I once sat in a room with a leadership team that had three options for a product launch. AI had modeled all three. Projected revenue, risk scores, timeline estimates. Option B was clearly the safest bet based on the data.

But I watched the room. Half the team was not bought in. The head of marketing had concerns she had not voiced yet. The engineering lead was doing mental math and his face said the timeline was fantasy. Everyone was waiting for someone else to speak up.

The project lead chose Option B because the data supported it. Six weeks later, they scrapped it. Not because the data was wrong. Because the people behind the plan were never aligned. The call nobody wanted to make was to slow down, name the tension in the room, and get honest alignment before committing.

AI gives you options. It does not make the hard call. It does not tell a senior stakeholder their pet project is not the priority. It does not say we are going to miss the deadline, here is why, here is how we recover. That is your job. And if you skip it, you are not leading. You are just forwarding recommendations.

4. Put your name on it.

I review a lot of deliverables. Strategy decks, project charters, stakeholder updates. And lately, I can tell within thirty seconds whether a leader actually shaped the work or just prompted it.

It is not about whether AI helped. I want you to use AI. The question is whether you sat with the output long enough to make it yours. Did you rewrite the opening because the AI version was accurate but did not land the way you need it to for this audience? Did you cut the section that sounds impressive but adds nothing? Did you make it sound like you?

I had a student once who sent a proposal to a client. It was clean. Well-structured. Technically sound. The client called her and said this is great but it does not feel like it came from you. That feedback stung. But it was right. The work was polished and it was empty. She rewrote it in her own voice, added her perspective, cut the fluff, and the client signed the next day.

Your credibility is built on the work that has your fingerprints on it. Not just your name at the top.

5. Deliver it in person.

This is the one I feel the strongest about because I have seen it play out so many times.

Early in my career, I spent a full week building a recovery plan for a project that had gone sideways. Scope changes, budget issues, a frustrated client. I put everything into that plan. It was solid. I emailed it to the steering committee and waited.

Nothing happened. No response. No feedback. I followed up. Got a polite reply about reviewing it next week. By the time it came up in a meeting, the moment had passed and half the room had already formed their own opinions without reading a single page.

The next time I had a recovery plan, I did not send it. I walked into the room. I presented it face to face. I answered every question in real time. I watched the faces when I laid out the hard parts. And when I could feel the CFO was not with me on the budget ask, I adjusted right there. That plan got approved the same day.

AI can build the deck. It cannot walk into the room and sell it. It cannot read the energy when the conversation shifts. It cannot stay after the meeting to have the sidebar that actually closes the deal. Delivery is not a detail. It is the whole game.

WHAT HAPPENS WHEN YOU DO THIS

I want to tell you what I have seen happen when leaders commit to closing the last mile. Because it is not theoretical. I have watched it play out.

I worked with a program manager last year who was drowning in deliverables. She was using AI for everything and her output was incredible. But her team was disengaged. Her stakeholders were polite but noncommittal. Things were technically getting done but nothing was landing.

We started applying the Last Mile Framework together. She began pressure-testing her plans before sending them out. She started adding her own context, the stuff only she knew from being in the room. She made two hard calls she had been putting off for weeks. She rewrote a strategy deck in her own voice. And she stopped emailing updates and started presenting them.

Within a month, her team felt different. They told her that. Not because the output changed. Because they could feel her in the work again. Her stakeholders started responding faster. Her plans started surviving first contact with reality because she had actually stress-tested them. And she got promoted six months later.

That is not a coincidence. That is the compounding effect of closing the last mile. In a world where every leader has the same AI stack, the ones who pull ahead will not be the ones with the best prompts. They will be the ones who are irreplaceably human in how they lead, decide, and show up.

THE REAL TEST

I want you to try something this week. Pull up the last deliverable you sent out. The one AI helped you build. Read it with fresh eyes and ask yourself one question.

Where am I in this?

Not your name. Not your title on the header. Where is your thinking? Where is the judgment call you made that the tool could not make? Where is the context you added from being the person in the room? Where is the hard truth you included because you had the courage to say it?

If you can point to those things, you are closing the last mile. You are using AI the way it should be used. As a starting line, not a finish line.

If you cannot find yourself in the work, that is not a failure. That is a signal. It means somewhere along the way, you started trusting the output more than your own judgment. And the fix is simple. Not easy. But simple.

Do the last 20%. Every time. No exceptions.

Because AI does not close the last mile. You do. And that is exactly why you still matter.

If this helped you, send it to someone who could use it this week.

Until next time,
Justin

✍️ From the Desk of Justin Bateh, PhD
Simple tactics. Real results. No fluff.