Cloud and Infrastructure Modernisation
Placement decided against data residency and jurisdiction first, with on-premises retained wherever the sovereignty position requires it rather than treated as failure.
Sectors ยท Government and Public Sector
Aging estates, split budgets and mandates that cannot slip. We sequence the work so that every phase stands on its own, which is what makes it defensible when the next tranche is late.
The pressure
What we are engaged to produce
A sequence where each phase is separately contracted, separately priced and independently defensible, so a deferred tranche costs momentum rather than the entire investment.
Government estates carry a particular combination we do not see elsewhere. The infrastructure is old, often significantly past end of support, and the funding to address it arrives in annual tranches that are approved separately and can be reduced or deferred without notice.
The usual consequence is a programme designed as one large transformation, part-funded, then stalled somewhere in the middle. The agency ends up carrying the disruption of a migration without the benefit of having completed one, and the next funding round has to be argued for from a weaker position.
There is also a sovereignty constraint that is genuinely binding rather than advisory. Where citizen data sits, who can reach it and under which jurisdiction are design inputs, not policy paragraphs added afterwards.
The instrument
Public programmes are funded in tranches and tranches slip. Every phase we plan has to leave the estate defensibly better on its own. Select a phase to see the position it ends on.
Even if nothing follows, the agency holds an accurate inventory and a ranked exposure register it did not have before, which is usually what the next funding round is argued from.
Each phase is contracted separately and priced separately. Nothing in a later phase is a precondition for the value of an earlier one, which is the only structure that survives a funding cycle honestly.
Applied here
Placement decided against data residency and jurisdiction first, with on-premises retained wherever the sovereignty position requires it rather than treated as failure.
Recovery on Veeam and VMware with retention set against statutory obligations, and restores tested rather than assumed across departments that rarely talk.
Energy audits that produce a defensible efficiency case, which in the public sector is frequently what unlocks the funding for the rest of the programme.
An end-of-support register that can be taken to a funding committee, ranked by service impact and costed, rather than a technical list nobody outside IT can act on.
A typical first engagement
A verified inventory with end-of-support positions, ranked by impact on the services citizens actually use.
Each phase separately priced and separately justified, structured so approval of one does not commit the next.
The same material rendered for a funding committee, in service and risk terms rather than in platform names.
Sequenced properly, a deferred tranche costs momentum. Sequenced badly, it costs the whole investment.