The Planning Illusion
It is November, and every enterprise technology leader is going through the same ritual: building a technology roadmap for 2026. They are estimating budgets, prioritizing initiatives, allocating headcount, and creating timelines that stretch twelve months into the future. And most of what they plan will be wrong by March.
Annual technology planning has always been an exercise in controlled fiction. But AI has made the gap between the plan and reality wider than it has ever been.
Why AI Breaks Annual Planning
- Capability shifts are unpredictable. No one planning their 2025 technology roadmap in November 2024 accurately predicted which AI capabilities would be available, at what cost, and at what quality level twelve months later. Planning your 2026 roadmap today faces the same problem. A model release in February could make half your planned AI investments unnecessary or could open possibilities you have not even considered yet.
- Build vs buy decisions are unstable. The decision to build a capability internally versus buying it from a vendor can flip in a single quarter as new products launch and pricing changes. A decision made in November may be obsolete by the time engineering actually starts working on it in February.
- Competitive dynamics are accelerating. Your competitors are also planning their AI strategies right now. If they ship before you, your planned differentiation may evaporate. If they stall, opportunities you deprioritized may become urgent. The competitive landscape in AI moves faster than annual planning cycles can accommodate.
A Better Approach
Replace the annual technology plan with a rolling quarterly strategy that includes three elements.
First, a fixed horizon for committed investments. Only plan in detail for the next 90 days. These are funded, staffed, and in execution. Do not change them without a significant reason.
Second, a flexible horizon for evaluated options. Maintain a ranked backlog of the next quarter's potential investments. Evaluate them quarterly against current capabilities, market conditions, and competitive dynamics. Commit to them only when they enter the 90-day execution window.
Third, a strategic direction that is stable. The overall direction, what markets you serve, what problems you solve, what differentiates you, should be stable year over year. Only the technology choices and implementation timeline should flex with the market.
The Political Reality
CFOs want annual budgets. Boards want annual plans. We understand the organizational reality. But within those constraints, build flexibility into the plan. Reserve 20-30% of the AI budget as unallocated, to be deployed against opportunities that emerge during the year. Label it explicitly as strategic reserve for AI optionality. It is a better use of budget than locking it into a plan that will not survive contact with reality.