All resources AI Strategy for the C-Suite

The Five-Year AI Plan Most Companies Should Refuse

The five-year AI plan is the wrong artifact for 2026. The 12/36/quarterly stack is what works.

TL;DR

The five-year AI plan is the wrong artifact for 2026. The right artifacts:

  1. A 12-month operating plan with named bets and concrete success criteria.
  2. A 36-month investment frame with capital allocation across the portfolio.
  3. A quarterly strategy review that adjusts the plan as capability changes.

The five-year plan looks confident and is wrong. The 12/36/quarterly stack looks less impressive and is correct.

The five-year AI plan is the wrong artifact. The 12/36/quarterly stack is what works.

The board asks for the five-year AI plan. The team produces it. Six months later, the plan is wrong because GPT-6 / Claude 5 / something else changed the unit economics. The team produces another five-year plan; another six months go by. The five-year plan is a ritual that produces no decisions. This piece is what to do instead.

Why the five-year plan fails

Three reasons.

1. Capability changes too fast. The unit economics of an AI agent in 2026 are 10–100x better than in 2024. Five-year plans built on 2024 economics were wrong by 2026. Five-year plans built on 2026 economics will be wrong by 2028.

2. Bet outcomes are unknown. The portfolio framing requires assuming most bets will fail. Five-year plans typically don’t accommodate this — they project all bets paying out. Reality is more like venture: 20% of bets produce 80% of value.

3. The artifact wrong-shapes decisions. A five-year plan is a commitment device; AI strategy needs to be a portfolio with options to kill, expand, or pivot. The wrong artifact for the actual work.

What to produce instead

The 12-month operating plan

Concrete. Named bets. Specific success criteria. Reviewed quarterly. Updated as needed.

Contents:

  • Named wedges (2–3 bets the company is making).
  • Success criteria for each.
  • Investment for each (people + dollars).
  • Timeline and milestones.
  • Operating governance (who decides what, when).

This is what the team executes against. Concrete enough to drive daily decisions; flexible enough to adjust quarterly.

The 36-month investment frame

Less concrete. Capital allocation. The shape of the portfolio.

Contents:

  • Total AI investment as % of revenue.
  • Allocation across the three categories (operational efficiency, product capability, business model bets).
  • Major infrastructure investments (platform, governance).
  • Talent strategy (build, hire, partner).
  • Risk and exposure budget.

This frames the 12-month plan. Updated annually with quarterly reviews.

The quarterly strategy review

The mechanism that keeps the 12-month plan alive.

Cadence: 90-minute review every quarter, with executive leadership.

Agenda:

  • Wedge progress: where is each bet?
  • Capability changes: what changed in the AI landscape?
  • New evidence: what have we learned that affects the plan?
  • Decisions: kills, reinvestments, new bets.

The output of each review is a updated 12-month plan and a refreshed view of the 36-month frame.

What boards actually need

Three artifacts for the board level.

1. The 36-month investment frame. Capital allocation, risk posture, governance status. Annual review with quarterly updates.

2. Wedge progress. What’s working, what’s not, what’s been killed. Quarterly.

3. Risk and exposure. Incidents, regulatory developments, competitive moves. Quarterly.

This is enough for board governance. The five-year plan adds nothing.

Counter: don’t shareholders / analysts want long-range plans?

For public companies, communication to shareholders matters. The right artifact for that audience:

  • A vision document (qualitative): where AI is taking the company, the strategic frame.
  • The 36-month investment frame (financial): how capital is allocated.
  • The named wedges (concrete): what specifically the company is doing.

These together communicate strategy without committing to fictional five-year specifics.

What about industries that need long-range planning?

Some industries genuinely need 5–10 year capital plans (aerospace, infrastructure, certain regulated sectors). For these:

  • The long-range plan covers infrastructure and organizational capacity.
  • The 12/36/quarterly stack covers the AI capabilities themselves.
  • Don’t try to put AI capability decisions into the long-range plan.

The infrastructure horizon and the AI capability horizon are different. Plan separately.

Common board reactions

Three you should expect.

1. “We need more confidence and detail.” The board wants the five-year plan because it feels accountable. The response: explain why the five-year plan is fictional confidence. Show the alternative.

2. “We need to commit to specific outcomes.” Specific outcomes are good for the 12-month plan. For five-year outcomes, ranges and scenarios — not commitments.

3. “Our investors expect a longer-range view.” A vision document covers this. The five-year plan does not necessarily satisfy investors and definitely doesn’t drive execution.

The board education is part of the work. Don’t skip it.

What to do this quarter

  1. Audit your current AI plan. Is it a five-year plan that’s already wrong? Common.
  2. Build the 12-month operating plan. Named bets, criteria, governance.
  3. Frame the 36-month investment. Capital allocation, not a roadmap.
  4. Schedule the quarterly review. Make it a real working session, not a status update.

FAQ

What if our planning culture demands five-year plans? Reframe. Provide the 36-month investment frame as the “five-year strategic context.” Provide the 12-month plan as “the year-one execution.” Provide quarterly reviews as “ongoing governance.” The artifacts are different; the rhythm satisfies the cultural need.

What about budget cycles that span multiple years? Multi-year budget cycles can accommodate multi-year capital allocation without requiring five-year capability plans. Budget the platform and governance over 3 years; plan capabilities over 12 months.

Doesn’t this make the company seem strategically vague? Strategically vague is “we’re investing in AI.” Strategically specific is “we’re betting on these three wedges, with this much capital, expecting these outcomes by this date.” The wedge-and-portfolio framing is more specific than most five-year plans, not less.

What about acquisitions? M&A in AI is mostly opportunistic — specific capabilities or talent acquired as opportunities arise. Build a target profile (what we’d buy), not a five-year acquisition plan.

How does this change for very large or very small companies? Same principle, different scale. Smaller: less formality, same rhythm. Larger: more formality (multiple BUs, more governance), same rhythm.

Working with JAIN on AI strategy artifacts? We help executive teams build the 12/36/quarterly stack that produces decisions instead of documents. Book a 30-minute call.

Related reading:

Want to talk through this for your team?

30 minutes, no slides. We'll work the specific call your company is facing.