LLenny's Podcast
← All frameworks
StrategyAsha Sharma (CVP of AI Platform at Microsoft)

Planning in Seasons

Replace rigid roadmaps with secular 'seasons', loose quarterly OKRs, and deliberate slack

Difficulty
Moderate
Time to result
~ongoing to results
Steps
4
Confidence
92%

A planning cadence for environments changing too fast for six-month roadmaps. Ground everyone in the current 'season' (defined by secular industry/customer shifts), set loose quarterly OKRs beneath it, run squads on 4-6 week goals, and deliberately leave slack in the system for both the unplanned and the 'slope' of future disruption.

Origin

Asha Sharma's adaptation of Microsoft's traditional six-month 'semester' planning, evolved for the pace of AI where a model like GPT-5 can drop mid-cycle and invalidate a roadmap.

Core principles

  • 01A season is defined by secular changes in the industry or from customers, not by a fixed calendar
  • 02Ground everyone on the ethos: what's changing, what customer problems to solve, what winning looks like, and the northstar metric
  • 03Seasons are uncomfortable and variable — they can last three months, six months, or a year
  • 04Leave slack not just for the unplanned but for the slope — how you'll disrupt your own platform
  • 05Openness to constant change is the business you're in, not a failure of planning

How to run it

  1. 1

    Name the current season

    Define the season by the set of secular changes happening in the industry or from customers (e.g. prototyping of AI → models and reasoning → the rise of agents). Align everyone on the ethos and northstar metric.

    Pro tip Give the season a shared name so the whole org centers its strategy on the same secular shift.

    Watch out Don't force a season onto a fixed six-month calendar — its length is dictated by the change, and could be 3, 6, or 12 months.

  2. 2

    Set loose quarterly OKRs

    Ask what you need to do next quarter to put yourself on a path toward the season's goal, and set OKRs loosely enough to survive mid-cycle surprises.

  3. 3

    Run squads on 4-6 week goals

    Have teams operate in squads with four-to-six-week goals for specific problem areas that ladder up to the quarterly OKRs.

  4. 4

    Leave deliberate slack in the system

    Reserve capacity not only for unplanned events but for the 'slope' — continuously investing in how you'll disrupt your own platform.

    Pro tip Treat slack as a permanent line item; it's what lets you absorb a GPT-5-style shock without derailing the whole plan.

In the wild

Microsoft AI platform season progression

Sharma's teams moved through seasons — prototyping of AI and early GPT work, then models and reasoning models, and now the rise of agents — re-grounding the org on each secular shift rather than a fixed roadmap.

Teams stay aligned on the current secular change while absorbing frequent disruptions, currently operating in the 'rise of agents' season with 15,000 customers deploying agents.

Common mistakes

Over-baking a six-month roadmap

Committing to a detailed half-year plan when the frontier can shift underneath you mid-cycle, leaving the plan stale and rigid.

No slack for the slope

Planning every unit of capacity against known work leaves nothing to invest in self-disruption or to absorb the unplanned.

Is it for you?

Best for

Product and platform leaders in fast-moving categories (especially AI) who need alignment without a brittle long-range roadmap

Not ideal for

Stable, slow-changing domains where predictable multi-quarter roadmaps still hold and seasons would add needless churn

From the transcript

we kind of think about it as you know what season are we in? And so a season which is very uncomfortable can be denoted…

31:00

we have kind of loose quarterly OKRs. So like okay if we believe that what do we need to do next quarter

32:00

teams are operating in squads and they're kind of setting out you know four to six week um goals

32:00

we try to leave slack in the system not just for the unplanned but for the the slope

32:30

From the episode

How 80,000 companies build with AI: products as organisms, the death of org charts, and why agents will outnumber employees by 2026

Asha Sharma (CVP of AI Platform at Microsoft)