Skip to main content

Insights

  • leadership
  • operations
  • ai
  • federal
  • culture

Grey Beards Last. Juniors First.

An innovative company lets the juniors speak first. The leaders guide, not squash. Here is the four-line meeting agenda we use to make that the default at Accelera Solutions, and why it matters more in AI work than it ever has before.

Look, I’ll say it before someone else does. I’m a grey beard. The last thing this organization needs in any of our rooms is more of me.

Here’s the other thing I’ve figured out over the years (back from my days running tugboats). A good deck crew can make any captain look great. A captain who never listens to their deck crew can put themselves, the vessel, the cargo, and the crew in jeopardy.

That lesson stuck with me long before boardrooms, proposals, or AI strategy sessions. When I speak last, the rest of the room speaks first. That’s where the work actually happens.

The default kills innovation

Most rooms run the wrong way. The senior speaks first; the room confirms. Whatever the senior said becomes the compass heading before anyone closer to the work opens their mouth.

That’s not a meeting. That’s verification.

I’ve watched this happen for years on both sides of the table. In a customer conference room it’s even worse, because now you’ve got two “senior-speaks-first” cultures stacked on top of each other. The customer’s senior speaks. Then ours speaks. The juniors on either side never really get a true turn. By the time the people closest to the problem finally open their mouths, the meeting is already over. Or worse, already committed to a course where innovation is no longer welcome.

An argument started with “I’ve been doing this for X years” is often an argument lost.

That’s something I say a lot. It usually lands wrong on the person who started the argument. But nobody in the room ever learned anything from how long someone had been doing the work. They learned something from whether the work was right.

Years aren’t an argument. Facts are.

What I do instead

I run our meetings on the same template every time. The agenda has four lines:

  1. The problem statement. What are we actually trying to solve? One sentence.
  2. What we know. Facts. Things we can point to. No opinions yet.
  3. What we don’t know. Honest gaps. Where we have to guess.
  4. What we need to succeed. People, time, decisions, information.

The order matters.

By the time we get to what we don’t know, the room is supposed to be uncomfortable. The people nearest the work, not the people highest on the org chart, are usually the ones with the answer.

One of the best lines from the book The Trusted Advisor:

The best way to sell yourself to others is first to sell others to themselves.

That’s leadership too. The job of the leader is not to dominate the conversation. It is to create the conditions where the people closest to the problem can think clearly enough to solve it.

Back on the tugboats, the captain in the wheelhouse who ignored a deckhand’s warning from the waterline was usually the captain everyone dreaded working for. The best captains knew when to step in, but more importantly, they knew when to listen, because the safety of the vessel, cargo, and crew depended on it.

Your juniors already know more than you do

Here’s the part where I get honest about how much I don’t know.

The world just spent the last few years inventing agentic AI. The people on my team who have been building with these models for two years know more about how they actually behave than I do. I’ve been doing federal and DoD IT a long time. None of that work taught me how to think about an agent that calls ten thousand APIs in an hour, or what a runaway token budget feels like in production at 2 a.m.

When I walk into a room about agent strategy, the smartest person in the room is rarely me. It is usually whichever junior engineer was awake at 2 a.m. the night before figuring out why the agent kept timing out.

If I speak first in that room, I slow us down.

If I let them speak first, I get to ask better questions. The organization gets to the right answer faster.

That’s how we work.

The grey beards in this organization, and I’m one of them, should talk last in the rooms where the work is actually getting done. The juniors talk first because they are closest to the problem. The leaders ask the question that sharpens the answer. And when the answer is ready, the leader’s job is to back it.

  • Dayel Ostraco

    Agents Are Just Identities. The Boundary Hasn't Moved.

    Federal IT already has the identity, authorization, lifecycle, and audit primitives agents need. The new work is governance at machine speed, not a new IAM tier. What the agent-platform market is mostly selling is an unnecessary layer.

     · 5 min read

Back to Insights