Home / Case Studies / Microsoft EA Renewal & Copilot
MicrosoftBankingEA RenewalCopilot

Microsoft EA cut by $9.6M, with Copilot ring-fenced.

A global banking group entered an EA renewal cycle with a proposed Microsoft 365 E5 step-up, a top-down Copilot commitment, and Azure pricing that no longer reflected the bank's actual consumption. Sixteen weeks later, the renewal landed at $9.6M below the opening proposal, with Copilot structured as an opt-in carve-out rather than a built-in commitment.

Enterprise corporate campus at sunrise
$9.6M
Three-year saving
22%
Reduction vs. opening proposal
16 wk
Kick-off to signature
0
Forced Copilot seats
The contract going in

EA renewal, Copilot pressure, Azure drift.

The bank's existing Enterprise Agreement was three years old and bundled Microsoft 365 E3, a small Power Platform footprint and an Azure commitment that had been sized against a now-obsolete migration plan. Internal demand for Copilot was real but uneven, sitting at a handful of pilot teams.

  • 40,000 M365 seats, mixed E3 and a small E5 pool.
  • Azure commitment $4.2M annually, with consumption tracking 60% against forecast.
  • Copilot pilot active for fewer than 800 users.
  • Renewal cycle running in parallel with a major core-banking modernisation.
Microsoft's opening position

E5 step-up, Copilot baseline, locked Azure.

Microsoft's proposal moved the full base to M365 E5, attached a fleet-wide Copilot commitment at preferred pricing, and rolled the under-consumed Azure commitment forward at a 9% uplift. The headline number was framed as growth-on-growth.

What we flagged

The Copilot baseline was being treated as a discount unlock for E5. Once committed, the floor would have been hard to reset on the next renewal, regardless of actual adoption.

The work

Sixteen weeks. Five workstreams.

1. Adoption evidence

We pulled real seat utilisation across the M365 stack and segmented the workforce into E3, E5 and lower-spec personas. The output: a defensible target mix that did not assume full-fleet E5.

2. Copilot opt-in structure

Designed a carve-out: agreed Copilot pricing locked at the EA level, but opt-in by business unit on a quarterly cadence, with no minimum commitment for years two and three.

3. Azure resize

The under-consumed commitment was rebuilt against an honest 36-month forecast tied to the core-banking modernisation milestones, with a Savings Plan layer for predictable workloads.

4. Security & compliance carve-outs

Worked with the bank's CISO to keep Defender and Purview where they made sense, without using them as a justification for fleet-wide E5.

5. Commercial position

We drafted the position paper. The bank's procurement lead presented it. Our team ran the support sessions across the four follow-up rounds.

Lesson

The opening Microsoft proposal usually bundles AI adoption into the discount mechanic. Separate the AI question from the EA question. They negotiate better apart than together.

The contract going out

Tiered seats, opt-in Copilot, right-sized Azure.

The renewed EA moved approximately one-third of the fleet to E5 against a justified persona model, held two-thirds on E3, and locked Copilot at a pre-agreed unit price available on quarterly opt-in. Azure was restructured against forecast with a Savings Plan layer.

$9.6M
Saved
Versus the opening Microsoft proposal, measured over three years.
Opt-in
Copilot structure
Quarterly opt-in by business unit. No fleet-wide floor.
36%
Azure rightsizing
Annual commit rebuilt against forecast, with Savings Plan layer.
“The advisor's most important contribution was separating the Copilot decision from the EA decision. Once we did that, the EA negotiated like an EA and the Copilot conversation could be paced against actual demand.”
Group Head of Procurement · Banking · Anonymised by client request
Related resources

More on Microsoft EA strategy.

← Back to all case studies

Have a Microsoft EA up for renewal?

Tell us the renewal date, seat count and Copilot position. We will respond within one business day with the lead and the most relevant precedent.