Engineering leaders spend an astonishing amount of time just finding things out.

Before every 1:1, there’s a ritual: check the team’s roadmaps, reference PRs, skim Jira and Jellyfish, scan Slack for anything that blew up overnight, read the last meeting notes to remember what was promised, cross-reference all of it against feedback and outstanding issues from other teams. Multiply that by eight direct reports, nine teams, 40+ ICs, and four peer groups. Before the first conversation of the day starts, hours are already gone – and most of it was just surfacing information that should have been waiting for me.

Then the meeting starts, and we spend half of it doing the same thing again. “How’s the sprint going?” “Any blockers?” “What happened with that incident?” We’re asking questions we could have already known the answer to, because there wasn’t time to dig through four systems before the calendar took over. All of this just to get to the work that actually matters.

This is the job many of us accepted. It felt like the job, and I don’t think it has to be.

What changed

About three months ago, I started building an AI system I call Sage – Strategic Advisory for Group Enablement, because I love a good acronym. It sits on top of the tools we already use: GitHub, Jira, Slack, meeting transcripts, engineering metrics. It synthesizes them into the context I actually need. Morning prep that tells me what shifted overnight. 1:1 briefings that surface open action items, coaching patterns, and things I should be asking about. Post-meeting reviews that tell me what we covered and what I missed. Health dashboards I can glance at instead of assembling from scratch.

I want to be clear about what this is and what it isn’t. Sage doesn’t make decisions. It doesn’t manage people. It doesn’t replace judgment or relationships or the hard conversations that only happen human-to-human. What it does is eliminate the archaeology – the hours spent excavating information from systems so I can show up genuinely prepared.

The mechanical details matter less than what happened next.

The time came back

When the information-gathering layer compressed, something shifted that I wasn’t fully expecting. My 1:1s changed. Not because the format changed, but because I stopped spending the first fifteen minutes catching up and started spending that time on things that actually matter: strategy, growth, coaching.

I started noticing patterns I’d been too busy to see. A manager whose team health looked fine on paper but whose coaching conversations had gone flat. Recognition opportunities that were slipping past because I didn’t have time to read every Slack channel. Decisions that were made three weeks ago and never revisited.

The shift isn’t “I have an AI that does my job”. The shift is “I have time to do the parts of my job that only I can do”.

We talk a lot about how engineering leaders should be coaches, should be strategic, should lift their heads and look around. The uncomfortable truth is that many of us don’t do enough of that – not because we don’t want to, but because the operational overhead of just staying informed eats the day. The information surfacing is the work, and it leaves surprisingly little room for the work that actually moves people and teams forward.

What I actually do differently now

Here’s a concrete example. Before Sage, prepping for a 1:1 with one of my managers meant checking their teams’ sprint boards, scanning relevant Slack channels, reviewing any open action items I could remember, and hoping I hadn’t missed something important. That’s twenty minutes minimum per person if I’m being quick, and I have eight directs.

Now I run a single command and get a briefing that includes all of that, plus coaching cues based on patterns from previous conversations. It flags when I’ve been solving problems instead of asking questions. It reminds me when someone committed to something three weeks ago and hasn’t mentioned it since. It tracks my own coaching habits and tells me where I’m falling short – it even surfaced that I show up as a better leader earlier in the day than later, which was humbling to see in data form.

That last part matters. One of the most valuable things Sage does is hold me accountable to the leader I’m trying to become, not just the operator I already am. It noticed – because it reviews my meeting transcripts – that I have a consistent pattern of jumping to solutions when a direct report is venting, instead of asking what they need from me. That’s the kind of feedback we don’t get from dashboards. We barely get it from coaches, because a coach isn’t in every conversation.

The real argument isn’t about AI

The thing I keep coming back to is this: the value isn’t the technology. The value is getting back the time and headspace that information-gathering was consuming, and redirecting it toward the work that actually develops people, shapes strategy, and builds culture.

Too many of us are competent operators who occasionally find time to coach. That ratio should be inverted. When the operational awareness moves to the background, coaching and strategic thinking become the primary activity – and that’s not a small shift. It’s the difference between a leader who’s always catching up and one who’s actually present.

I acknowledge more work, I catch more patterns, I follow up on commitments, I prepare for skip-levels instead of winging them. I have time to think about problems that aren’t on fire. These aren’t things I couldn’t do before; they’re things I didn’t have consistent time for.

One of my direct reports recently told me that walking into 1:1s already knowing what they’d worked on that week makes their work feel seen. They don’t have to “sell” or “report” what they did anymore. That landed harder than any metric I could point to.

You should build your own

Here’s the part where it would be easy to point at a product and say “go buy this.” I’m not going to do that, and the reason is important.

The most valuable thing about Sage isn’t the synthesis or the dashboards. It’s that it’s shaped around the specific gaps in how I lead. I know where my weaknesses are. I know which parts of the job I tend to avoid, which patterns I fall into under pressure, which information I consistently fail to surface on my own – especially when I’m feeling time-poor. Sage is built around those gaps, not around a generic “engineering leader” template.

That means building your own version, even a simple one, is more valuable than adopting someone else’s. You don’t need what I need. Maybe you’re great at coaching but terrible at following up on decisions. Maybe you never miss a recognition opportunity but consistently avoid difficult feedback conversations. Maybe your problem isn’t 1:1 prep; it’s that you haven’t looked at your team’s actual output in weeks because the calendar won’t let you.

The tools to do this are available right now. Claude, GPT, Gemini – pick one. Connect it to your existing systems. Start with the thing you hate doing most, or the thing you know you’re worst at. It doesn’t need to be sophisticated. My first version was a single prompt that summarized meeting transcripts, logged decisions, and recorded tasks. It grew from there because every time it saved me thirty minutes, I could see the next thing it should do.

I should warn you: this can become addictive. Until recently, most of the time I was saving went right back into building the system. Sage actually suggested at one point that I should focus on using it rather than continuing to tinker. As my favourite manager once said, “Why have a dog and bark yourself?” It pays to listen to your creation.

The building is also the learning. Every hour I spend shaping Sage teaches me something about AI that I can bring to my teams, my org, and the technical conversations I need to be credible in. Engineering leaders who aren’t building with AI right now are accumulating a gap that will be very visible very soon. Not because AI will replace us, but because the leaders who use it will simply operate at a different level: more informed, more present, more strategic.

The question isn’t whether AI can help you lead

It can. The question is what you’d do with three hours back every day. What would you pay attention to if you weren’t buried in information retrieval? Who would you coach if you had time to prepare? What patterns would you notice if you could actually look?

Most of us already know the answer. We’ve just been too busy finding things out to act on it.