Everyone Agreed on the Outcome. Nobody Would Sign for the Path.

"I Cannot Commit a Date I Cannot Defend"

Charter Milestone 2 required a committed roadmap by 15 April 2025. That was seven weeks away. The program manager had booked a two-hour working session with Dmitri Volkov, the platform integrator delivery lead, to walk through the core banking integration schedule and land a date. Forty-five minutes in, the session had produced no dates.

The program manager's first instinct was to push anyway. Vendors often say they cannot commit when what they mean is they would prefer not to. But Volkov was not refusing. He was telling the PM something specific about the technical substance. The PM asked Priya Raman, technical lead, to join the conversation the next morning.

Raman's read, delivered in her own fifteen-minute working session on Day 46, made Volkov's position legitimate.

The PM had two options. Option one: push Volkov to commit a date with a substantial risk reserve, and accept that the reserve would hide the uncertainty from the roadmap. Option two: make the uncertainty visible on the roadmap itself, and commission the work required to reduce it before the 15 April commitment date.

The first option was faster. The second was honest.

Three Levels. Not One.

The program manager went to Ahmed Hassan with a proposal on Day 47. Rather than publishing a roadmap with single dates that hid uncertainty, the Falcon roadmap would carry a Commitment Level field for every milestone. Three levels, labelled plainly so no stakeholder could mistake one for another.

Every milestone on the roadmap would carry one of these labels, visible to every stakeholder, with the assumptions that qualified a Proposed or Indicative date written next to it. A stakeholder reading the roadmap would know exactly which dates they could plan against, which they should treat as intent, and which they should treat as targets.

The Volkov question now had a clean home. The core banking integration milestone could be recorded as Proposed, with an explicit assumption: “Subject to core banking query performance within 20 percent of profiled baseline, data profiling work stream completing by 14 March 2025.” Volkov could sign that. The roadmap would carry it as Proposed until profiling data was in, then convert to Committed at the next steering if the baseline held.

The second half of the solution was commissioning the profiling work stream itself.

By the time the roadmap went to steering on 28 March, profiling had completed. The core performed within expected bounds at baseline load. Volkov signed. The integration milestone converted to Committed at the 3 April steering, two weeks before the charter's 15 April deadline.

The Pull-Forward That Never Reached the Room

Samuel Osei had learned from kickoff. The direct scope request (Post 04, USSD-later) had not worked. On 18 March, three days before the roadmap workshop, Osei sent the program manager a note in Slack: “Worth checking if Market A can pull forward three weeks. Retail's Q2 board story needs an anchor date. Happy to discuss in the workshop.”

This was the out-of-cycle protocol doing its work. The Engagement Plan (Post 04) routed scope-adjacent requests to PMO written response, not to workshop agendas. Hassan saw the message, flagged it to the PM the same afternoon.

The pull-forward request was not unreasonable in isolation. Retail's Q2 board cycle ran mid-June. An anchor date in the Falcon roadmap would be valuable to Osei's own governance. But pulling Market A forward three weeks meant compressing UAT, which was the go-live gate Okonkwo held veto on, which meant the central bank certification window shrank from six weeks to three. That trade was not Osei's to make alone. It was a steering decision.

The program manager made a seven-minute phone call to Fatima Idris on the evening of 19 March. The call did three things: surfaced Osei's request, explained the UAT-compression consequence, proposed handling it upstream rather than in workshop.

Idris agreed. The Idris-Osei-PM call happened on 20 March, thirty-five minutes. Osei understood the trade immediately, asked for a quarterly retail-specific readout he could use in his own board cycle (granted, added to the comms plan), and dropped the pull-forward request. The roadmap workshop on 21 March did not relitigate Market A dates.

Drafting the Roadmap

Roadmap Draft, Extract

A representative extract of the AI output, showing the commitment-level discipline at work. The full output covered 14 milestones across the four phases; the six shown below are where the commitment levels differed meaningfully and where the flagged items landed.

Program Roadmap, Extract

Extract of the published Program Roadmap v1.0, 28 March 2025. Six milestones shown from a total of fourteen. Every milestone carries its commitment level as a first-class field.