How I Use GenAI as a Thought Partner, Not a Shortcut

You don’t need to be a power user to get powerful results. I’m not training models or prompting GPTs into poetry—I’m just using them to do what great managers already try to do: communicate clearly, prioritize outcomes, and lead with intention.

Over the last few quarters, I’ve built a handful of custom GPTs to support my weekly, monthly, and quarterly workflows. These tools help me beat blank-page syndrome better than anything else I’ve tried, and they get me through the ugly first draft fast. Here’s how I’ve built AI into my rhythm as an engineering leader.

Weekly updates that speak to impact, not just output

Fridays used to be consumed by writing updates: compiling Slack threads, tickets, and documents to distill a narrative from dozens of half-formed notes. I wanted the update to show what we accomplished, and with lots of meeting distractions, it was easy to run out of steam and think, “I’ll just publish on Monday."

Now, I use a custom GPT that interviews me like a coach or a chief of staff:

• What shipped?
• What are we proud of and/or worried about?
• Any customer quotes, metrics, or open questions?

I drop in raw notes and thoughts, and it asks follow-ups, like, “Why does this matter?” or “What outcome does this enable?” It nudges me to think beyond the activity to the impact of our work. It even formats the output in Slack-ready bullets, complete with emoji ✨. I still spend time gathering notes, but the mental energy I save in editing makes the whole process lighter. Now, our work shows up with the context and clarity it deserves, and leadership stays informed without asking. This is how I communicate up and out without losing half my Friday.

Monthly reviews that keep strategy in focus

Every month, I pause to zoom out: Are we doing the right things, or just doing things? Are we investing where it matters?

I used to have a calendar invite with a few reflection prompts, but it was really easy to ignore. It was easy to put off the important but non-urgent work of reflection when the prompts began to feel stale, and my notebook reminded me of the to-dos on my list that felt more urgent. Just like with coaching, a conversation gets me unstuck in a way that solo reflection does not. Now I use gen AI to get the reflection started.

Personal leadership reflection:

• Am I building the career I want?
• Am I creating growth for others?
• Am I asking for help when I need it?

Team and portfolio review:

• Which projects are delivering business impact?
• Which ones aren’t? And why?
• Are there projects that need a shift in ownership, scope, or support?

It also prompts escalation points and spotlights misalignments before they turn into drift. This monthly reflection feels lightweight, which means I can actually reap the benefits of the practice.

Quarterly business reviews that practically write themselves

At Honeycomb, QBRs are a chance to look back on the last quarter and tell the full story: what we shipped, how systems performed, what bets we made, and what’s next.

Instead of starting from a blank doc, or even the QBR template, I use a GPT to co-write with me. It guides the process section by section:

• Product health and metrics
• Customer-facing milestones
• Strategic bets
• SLOs
• Risks and dependencies

If I decide to skip a section (for example, I often like to wait on SLO results until the last possible moment), it holds space for it so nothing falls through. Then it pulls it all together into a clean, narrative-style summary with trends highlighted, gaps surfaced, and a draft executive summary written based on what we built together. I no longer struggle with a single, large, daunting task of writing the QBR. Instead, I move section by section, a little at a time. I trust the GPT to remember my context and help me pick up where I left off. In the end, I have a complete, thoughtful QBR with a clear story of customer value and team momentum.

Where generative AI can’t replace a manager’s work

I don’t outsource judgment, expectations setting, or anything else that requires nuances or an understanding of how humans behave in different contexts. One clear example is managing performance. I’ll use GPT to help draft performance reviews, but will never rely on it to evaluate performance.

The real work of performance management isn’t just about summarizing projects or measuring outcomes. It’s about understanding growth over time, surfacing hidden effort, discerning contribution, and interpreting subtleties that won’t show up in documentation. AI can’t see the quiet mentoring, the extra effort, or that there is one who always raises their hand, or raises the bar. Generative AI can help me find the right words, but it can’t really see how each person’s contribution makes the whole team stronger, or how the group ends up being more than just the sum of its parts.

You don’t need to be a power user. You just need a prompt.

The custom GPTs I’ve built aren’t the most sophisticated use of AI. They’re structured conversations that make communication better. They help me think more clearly, reflect more often, and write more intentionally.

If you’re an engineering leader, I’d encourage you to build your own. Pick a recurring task. Be extremely clear on how it can serve you best, the tone that works for you, and the audience you are writing for. Build a GPT that acts like a coach or a thoughtful peer and let it guide you to deeper reflection and more intentional communication.