The Copilot Credit Pre-Purchase Plan (P3) lets enterprises buy an annual pool of Copilot Credit Commit Units (CCCUs) at discounted pricing — 5% off for small commits, scaling to 20% off for the largest. The pool runs for one year (not monthly), addressing the main limitation of Capacity Packs. Sizing the commit correctly is the central decision: at 70–80% of expected annual consumption, the economics work cleanly. At 100%+ sizing, overcommitment risk dominates. At under 50%, undersizing forces expensive pay-as-you-go for the balance. The 70–80% sweet spot delivers 8–15% effective savings on total annual Copilot Studio spend for most enterprises.
The discount tier structure
The pre-purchase discount scales with commit size. Specific tier breakpoints are not publicly fixed — Microsoft account teams have commercial discretion on tier placement — but typical pattern:
| Tier | Approximate annual commit size | Discount |
|---|---|---|
| Entry | Modest enterprise volume | ~5% |
| Mid | Significant enterprise volume | ~10–12% |
| Strategic | Large enterprise volume | ~15–18% |
| Largest | Strategic Microsoft customers | ~20% |
The non-public tier breakpoints create a negotiation lever. Procurement teams whose commit sits at a tier boundary can push for the higher discount tier through commercial pressure or by stretching the commit modestly to clear the breakpoint. The 2–3 percentage point difference between tiers compounds substantially across a large commit.
CCCU sizing methodology
Right-sizing is the central commercial decision. Three components feed the sizing model.
Component 1: Baseline consumption
The expected steady-state monthly consumption based on current usage patterns. Ideally derived from 3–6 months of pay-as-you-go data establishing actual consumption rather than projected consumption. Forecast-only sizing is risky because Copilot Studio consumption forecasts have historically been inaccurate by 30–50% in either direction.
Component 2: Growth trajectory
Expected consumption growth over the commit year. Most enterprise deployments grow consumption through the year as new agents reach production and existing agents expand usage. Modest growth assumption (10–30% over baseline) is typical and conservative.
Component 3: Variance buffer
Buffer for month-to-month variance. Even predictable workloads have variability driven by business cycles, project timing, and adoption pace. A 10–15% variance buffer prevents normal variability from forcing pay-as-you-go overage.
The combined sizing target: monthly baseline × 12 × (1 + growth assumption) × 0.85 = annual CCCU commit. The 0.85 factor sizes the commit at 85% of full-coverage expected demand, intentionally leaving 15% spillover to pay-as-you-go. This protects against overcommitment while capturing the discount on the predictable portion.
The two sizing failure modes
Failure 1: Overcommitting
Sizing the commit at 100%+ of expected consumption. Unused CCCUs at year end are lost — they do not roll forward and cannot be transferred. Common cause: optimistic growth projections that do not materialise. Worst-case loss: 20–40% of the commit if consumption substantially underperforms expectations. The right protection is conservative sizing rather than optimistic forecasting.
Failure 2: Undercommitting
Sizing the commit at 50% or less of actual consumption. Pool exhausts mid-year, forcing pay-as-you-go premium for the remaining months. The math: a 15% discount on the committed half is offset by 0% discount (full pay-as-you-go rate) on the uncommitted half, producing roughly 7.5% effective annual discount — significantly below what proper sizing delivers. Common cause: insufficient baseline data or excessive conservatism.
Multi-year CCCU commits
CCCU pre-purchase defaults to annual commitment. Multi-year (2–3 year) commits are negotiable for strategic customers and capture additional discount on top of the volume tier — typically 1–3 percentage points per additional year locked. A three-year commit at strategic tier can therefore reach an effective 22–25% discount versus pay-as-you-go.
Multi-year commits have an obvious risk: consumption forecasts deteriorate over longer horizons. The right protection is contractual flexibility — including reset rights, downward adjustment provisions, and explicit treatment of consumption variability. Microsoft will accept these provisions for strategic customers; the procurement work is explicitly negotiating them.
Three negotiation levers
Lever 1: Tier placement. Push for the higher discount tier on borderline commits. Microsoft account teams have discretion and exercise it for prepared procurement teams.
Lever 2: Multi-year commitment with reset rights. Two- or three-year commits at higher discount with explicit annual reset rights protect against forecast deterioration while capturing additional commit discount.
Lever 3: EA bundling. CCCU commits bundled into EA renewals provide a single procurement vehicle and additional leverage. The bundling helps account teams meet attach quotas, which can deliver flow-back discount on other EA line items.
Action plan for CCCU evaluation
- Gather consumption baseline data. Three to six months of pay-as-you-go data ideally; otherwise rigorous forecast methodology.
- Size the commit at 80–85% of expected demand. Intentional underweighting protects against overcommitment.
- Confirm tier placement and negotiate up if at boundary. 2–3 percentage points is material at scale.
- Consider multi-year commit with reset rights. For strategic customers with stable forecasts.
- Engage independent advisory. The sizing decision is genuinely material and benefits from external pressure-testing. Book a scoping call.