Year-End Retro: What Worked in 2024

I run retrospectives for my teams constantly. It is only fair to run one on myself. Here is what worked for me in 2024 and what I am carrying into 2025.

What Worked

The capacity planning framework. Building a standardized capacity model early in the year was the single highest-ROI investment of my time. It saved me hours of ad-hoc spreadsheet work every sprint and eventually became the org-wide standard. Solving your own problem well enough that others want your solution is the best kind of influence.

AI as a thinking partner. I moved from using AI for drafting text to using it for reasoning about complex problems. o1-preview for scenario modeling, Claude 3.5 Sonnet for code review discussions with engineers, GPT-4o for quick formatting tasks. This shift made me noticeably more productive in Q3 and Q4.

Saying no to meetings. I declined approximately 30% of meeting invitations this year. Not rudely — I asked for the agenda and suggested async alternatives when the meeting had no clear decision to make. This recovered about five hours per week. Those hours went into actual program work.

Investing in documentation. I documented every repeatable process I use — change management, risk triage, budget review. When I took on my third account, these playbooks meant I could onboard without starting from scratch. The investment paid for itself within a month.

What Did Not Work

Trying to be in every conversation. For the first half of the year, I attended too many meetings across all three accounts. I burned out by June. The fix was delegation — trusting team leads to handle operational decisions and briefing me asynchronously.

Delayed escalations. I absorbed two problems in Q2 that I should have escalated earlier. Both grew larger than they needed to. I recalibrated my escalation judgment in the second half.

Going Into 2025

I am focusing on two things: deepening my AI governance knowledge as it becomes central to enterprise delivery, and building repeatable playbooks for every program lifecycle stage. The goal is to operate at a level where taking on a fourth account is sustainable.