LLenny's Podcast
← All frameworks
StrategyFiona Fung (Manager of the Claude Code and Cowork Teams)

Just-in-Time Planning

Shrink long roadmaps to a lightweight monthly priority list and grant explicit permission to kill dead processes.

Difficulty
Easy
Time to result
~weeks to results
Steps
4
Confidence
88%

Because the field changes so fast, Fiona Fung replaced six-month roadmap docs with JIT (just-in-time) planning: a lightweight monthly priority list (a small spreadsheet, not docs) checked weekly to confirm the priorities still hold, over a backdrop of themes set roughly every six months. Underpinning it is a cultural rule — explicit permission to kill processes that no longer serve — so any dreaded, noisy, expensive, or manual process is regularly re-examined for whether it still has a purpose.

Origin

Fiona Fung, from her first big learning joining Claude Code when a lightweight six-month roadmap she introduced went unreferenced three months in.

Core principles

  • 01A fast-changing field makes long-horizon plans go stale before you reference them
  • 02Plan just in time: a lightweight monthly priority list, verified weekly
  • 03Keep long-range themes loose, set roughly every six months
  • 04Grant explicit permission to kill any process that no longer serves its purpose

How to run it

  1. 1

    Set loose six-month themes

    Bring the whole team together roughly every six months to agree on themes of where the work is heading, without heavy detail.

  2. 2

    Plan one month at a time, lightweight

    Align on the month's highest priorities in a simple spreadsheet — not docs — and let each person own how they address them.

    Pro tip Keep the list short and focused so updating it never feels like a tax.

  3. 3

    Re-confirm weekly

    Each week do a quick check that the listed priorities are still this month's priorities, adjusting as the landscape shifts.

  4. 4

    Audit and kill dead processes

    Periodically pick one process you dread, that's noisy, expensive, or very manual, and ask whether it still serves its purpose — kill it if not.

    Pro tip Look for ways to automate the surviving process so it stops feeling like a tax.

    Watch out Even a process you introduced yourself can outlive its purpose; be willing to cut your own.

In the wild

The unreferenced six-month roadmap

Joining Claude Code, Fiona introduced a deliberately lightweight six-month roadmap doc. Three months in she realized the team hadn't referenced it because so much had changed, so she cut it down to JIT monthly planning on a small spreadsheet.

Planning shrank to a monthly priority list checked weekly, eliminating wasted planning effort while keeping the team aligned in a fast-moving field.

Common mistakes

Writing long roadmaps in a fast-changing field

A six-month plan goes stale and unreferenced within months when the landscape shifts quickly, wasting the planning effort.

Keeping a process because you always have

Failing to ask whether a dreaded, noisy, or manual process still serves its purpose lets dead work persist; the fix is explicit permission to kill it.

Is it for you?

Best for

Leaders of fast-moving teams (especially in AI) who want alignment without heavyweight roadmap overhead

Not ideal for

Long-lead, capital-intensive, or heavily regulated programs that genuinely require multi-quarter committed plans

From the transcript

I I call it JIT planning now, like just-in-time planning.

1:24:30

we've shrunk it to JIT monthly planning.

1:25:00

explicit permission to kill processes that no longer serve us.

1:23:00

always be open to learning and always ask yourself whatever process you have is it still serving its purpose? Just because the field is changing…

1:24:00

From the episode

What happens after coding is solved?

Fiona Fung (Manager of the Claude Code and Cowork Teams)