Mike San Román's homepage

How I used Claude Code to accelerate strategy work

August 20, 2025

I had solid strategic work done: it had already gone through three iterations with feedback and strong alignment with my engineering leadership team at Buffer. I thought I had it, but Joel Gascoigne helped me see I needed to approach our engineering strategy from a completely different angle.

I faced a familiar challenge: I had all the strategic thinking I needed, but organizing it into a new structure felt like starting from square one.

That's when strategy can be the enemy of operations. When you can't quickly access and restructure your existing strategic insights, teams start waiting for direction that keeps shifting. Momentum dies.

Then I remembered this tweet from an Anthropic employee talking about how everyone there uses Claude Code for everything (not only coders!). And it got me thinking and excited about the possibilities.

I'd been wanting to migrate my 1,400+ notes from Reflect to Obsidian anyway. The idea of having direct access to all my notes as plain Markdown files always felt empowering, but the migration had always felt cumbersome. Maybe with Claude Code it could be different?

So I exported my Reflect data, fired up Claude Code, and said "Figure out how to migrate all this to Obsidian based on my organization patterns," before jumping into a meeting.

The magic moment

When I came back from the meeting, everything was migrated. Claude had written a script and just... did it. With my existing structure and format. Nothing was missing.

I had Claude process all the migrated notes relevant to the task at hand. Years of meeting transcripts, team feedback, strategy documents. Not to generate new ideas, but to help me access patterns across my own documented strategic thinking that I couldn't see while buried in the details.

What happened surprised me.

Claude helped me organize the strategic insights I had already developed into structures that were actually finishable. This only worked because of the solid strategic foundation we'd built through months of team discussions and documented experience.

My previous iterations were solid, but the themes could overlap, drag on forever, or leave you wondering "are we done?" Claude helped me restructure them into something completely different.

From endless to executable

The transformation wasn't about finding new strategic directions. It was about making my existing work executable by restructuring each theme to be:

  • Independent: No dependencies between themes
  • Ambitious: Real transformational outcomes, not incremental improvements
  • Solid: Clear "we'll know we've succeeded when" criteria
  • Clear: No ambiguity about what we're actually building
  • Checkoff-able: Definitive completion states

The original six detailed (but potentially endless) themes quickly transformed into five strategic pillars with clear finish lines.

For example, instead of "Simplify to scale" with examples that could go on forever, we now have "Make Buffer's tech stack boring" with a specific transformational outcome: "Single, unified technical foundation across all of Buffer."

We know we've succeeded when our legacy PHP to TypeScript migration is complete, our MongoDB setup is healthy and performant, frontend is consolidated to one app, and all teams use consistent patterns.

That's not a theme you iterate on indefinitely. That's work you complete and check off.

The broader insight

This experience taught me something important about strategic work in fast-moving organizations.

The problem often isn't that you need more strategic thinking or fresh perspectives. It's that you need faster ways to access and organize the strategic thinking you already have.

When you've done solid strategic work, the bottleneck becomes organizing that thinking into executable structures. The faster you can access your own insights, the faster you can act on them.

Claude Code became my decision compression tool. Not because it generated strategy for me, but because it compressed the time between having insights and structuring them for execution. If my underlying strategic foundation had been weak, organizing it faster wouldn't have made it better.

The technical details

For anyone curious about the actual migration process, here's what happened:

  1. Export: Reflect's export feature gave me a folder of Markdown files with all my notes
  2. Processing: Claude Code analyzed the file structure (and the file structure I had in a previous Obsidian test) and created a migration script that:
    • Organized notes into a logical folder hierarchy in Obsidian
    • Preserved all links and metadata
    • Created an index system for quick navigation
  3. Strategy synthesis: Once everything was in Obsidian, Claude could cross-reference meeting transcripts, feedback documents, and strategic thinking to identify patterns

The whole technical migration took maybe 20 minutes of actual work. The strategic breakthrough took a couple hours of conversation with Claude about how to structure completable outcomes.

What this means for engineering leadership

As engineering leaders, we're constantly balancing strategic work with operational effectiveness. The challenge isn't just making good strategic decisions: it's making strategic decisions that enable rather than block operational momentum.

Some lessons from this experience:

  • Build systems for rapid perspective shifts: When you need to reframe strategic problems, having tools that can quickly synthesize existing knowledge saves weeks of traditional iteration.
  • Structure work for completion: Strategic themes should have clear endpoints, not endless improvement cycles. If you can't check it off, it's not a theme: it's a process.
  • Use AI to accelerate existing expertise: Claude didn't replace strategic thinking: it helped me synthesize strategic insights my team and I had already developed. It compressed the time between having insights and structuring them into executable work.
  • Protect operational velocity: Strategy work that drags on too long becomes its own form of technical debt. Sometimes the best strategic decision is to make a decision and move forward.

By using AI to compress how quickly I could organize existing strategic thinking, I ended up with better outcomes. When you can rapidly access your own insights and structure them for completion, you naturally create clearer, more actionable plans.

Strategy shouldn't block operations. Build systems that help you quickly organize the strategic work you've already done, and watch how much faster your team can move.