A Year in Review: What Actually Worked Across Client Projects in 2026

Every year-end review I've read from other consultants is a list of trends and tools. Mine isn't, because looking back honestly at this year's projects, almost none of what moved outcomes was a tool. It was a short list of boring habits, repeated across very different engagements.

What actually worked

The single biggest driver of good outcomes this year was writing things down before starting, not after. Every project that went smoothly had a written scope, a written measurement plan, or a written success criterion agreed before implementation began. Every project that went sideways was missing at least one of those, and the sideways parts were almost always traceable back to that specific gap, not to any technical failure.

The projects that went well in 2026 weren't the ones with the best technical execution. They were the ones where the fewest things were left unwritten.

What I expected to matter and didn't

• Which AI agent framework a client's team used mattered far less than whether they'd built an eval harness before tuning anything — the framework choice was mostly noise.
• Server-side tagging platform choice mattered less than whether anyone was actually monitoring the container after launch.
• A fancier dashboard rarely outperformed a smaller one paired with a scheduled habit of actually looking at it.
• Client sophistication with AI tools mattered less than whether they had a specific, concrete use case already identified before engaging me.

The pattern underneath all of it

Every lesson this year traces back to the same root cause: ambiguity left unresolved early gets expensive later, and clarity invested early pays off disproportionately. That's true of scope documents, measurement plans, stakeholder alignment, and even the AI projects where the eval harness mattered more than the model choice — in every case, the unglamorous upfront work was the actual differentiator, not the sophistication of what got built on top of it.

Going into next year, that's the one thing I'm actually changing deliberately: spending more of the early project budget on making things explicit, even when a client is eager to skip straight to building. It's the least exciting part of the work and, based on this year's evidence, the part doing the most of the actual work.