Public Sector Cloud Waste in 2026: How One Owned System Replaces Dozens of SaaS Licences

Public sector SaaS cost reduction does not start with a new tool, it starts with the ones a department already pays for and barely uses. A central government team, a local council or an NHS back office can fold Power BI, Tableau, Salesforce, Dynamics and ServiceNow into one owned system that runs offline on its own hardware, with no per seat licence, no per ticket fee and no cloud egress. The money that leaks into idle seats and data charges stays inside a capability the organisation keeps, rather than renewing into another year of subscriptions.

Public sector SaaS cost reduction starts with the licences you already waste

The 2026 Spending Review told departments to use AI to take cost out of the back office, not to add another subscription on top of the ones they run. Industry reporting puts the UK public sector cloud market near 6bn pounds a year, and the Government Digital Service has been building a central data solution precisely because no single team could see how much of that spend was waste. Flexera's 2026 survey of cloud decision makers put wasted cloud spend at 29 percent, the first rise in five years. In the public sector that waste has a plain name: seats bought for staff who moved on, dashboards nobody opens, and a per ticket meter on a service desk that could deflect most routine work.

The pattern repeats across arms length bodies and shared service centres. Each function buys its own cloud subscription, each bills per seat, and each renewal creeps up whether the seats are used or not. Consolidation is the obvious answer, and now a directed one, but consolidating onto another cloud suite swaps one meter for another. The saving that holds is the one where the meter stops entirely.

What the sprawl costs a department today

Three costs stack up under the current model. First, the per seat multiplier: analytics, CRM and service desk each charge for every named user, so a headcount rise is a licence rise, and a reorganisation leaves orphaned seats that still bill. Second, the usage add ons: vendors that once charged per seat now charge again per conversation and per automated resolution, so the AI a department wants sits on top of the fee it already pays. Third, the cloud tax: hosting and data egress are charged per gigabyte leaving the region, and a governed public body pays that on top of the licence for the privilege of moving its own records. Around all of it sit the consultants a department hires to make the separate clouds talk to each other.

One owned system on hardware the department already runs

Mickai is one sovereign system that runs offline on the department's own hardware. It is built from studios, and a studio is a ready-made application for one business function that lives inside the single system rather than as a separate subscription. Mickai carries sixty-three studios in total, ten production-ready at launch and fifty-three in development, and the ones a public sector back office needs most are already in the launch set. Nothing calls out to a vendor cloud, so there is no per seat charge, no per conversation add on and no egress line, because no data leaves the building.

Omni, the Assistant, is the single front door: one prompt routed across the fifty brains on hardware you own. Staff ask in plain language and Omni sends the request to the right studio, so a department does not train people on five separate consoles. The Assistant runs on the organisation's own brain, built on its own data, and every AI action is sealed under post-quantum cryptography into a signed record. That record, the Open Audit Record, is what an accounting officer and an internal auditor can later stand behind.

The studios that replace the public sector stack

Three studios carry most of the back office load. Pythia, the analytics studio, answers questions in plain language over your governed on-premise data, from a single KPI to a full board-ready read, and stands in for Power BI, Tableau and ThoughtSpot with no analyst seat to license. Xenia, the customer relationship management studio, scores and prioritises contacts, drafts outreach and segments records for casework and citizen engagement, replacing Salesforce, HubSpot and Microsoft Dynamics. Ponos, the IT service desk studio, triages and routes a ticket, runs incident, problem, change and request against the configuration database and proposes a resolution from the knowledge base, replacing ServiceNow, Jira Service Management, BMC Helix and Freshservice.

Because the three sit inside one system and answer to one Assistant, the integration work that used to need consultants is simply how the studios already talk to each other. A service desk ticket, a CRM record and an analytics read draw on the same governed store, on the same hardware, sealed to the same record.

What you replace, and what you save

The audit trail an accounting officer can stand behind

Public spending has to be defensible, and so does public AI. Every action inside Mickai, an analytics read, a CRM change, a service desk resolution, is sealed under post-quantum cryptography into the Open Audit Record, a signed timeline that shows what the system did, on what data and when. That evidence supports SOC 2, ISO 27001 and GDPR examinations rather than claiming the certificate is already held, and it gives an internal auditor a record that never left the department to interrogate. For a body that has to answer to Parliament, a regulator or the public, the trail being on owned hardware is the point, not an afterthought.

How a department consolidates

The move from a cloud stack to one owned system runs as a short, ordered sequence rather than a rip and replace.

• Inventory the current subscriptions and the seats actually in use, so the idle spend is visible before anything is signed away.
• Stand up Mickai on the department's own hardware and point it at the governed data store the records already sit in.
• Bring Pythia live first for analytics, so the dashboards staff rely on are reproduced with no per seat licence.
• Move casework and engagement into Xenia, and route the service desk through Ponos, retiring the CRM and ITSM seats as they are replaced.
• Let the subscriptions lapse at renewal rather than mid term, converting recurring cloud opex into one owned capability, with every step sealed to the Open Audit Record.

None of this trades capability for savings. The analytics, the casework and the service desk are the same functions, run by the same teams, on the same records. What changes is where the money goes: into hardware the organisation owns, not seats and egress it rents and wastes.

Frequently asked questions

How does a sovereign system deliver public sector SaaS cost reduction?

It removes the meters. Analytics, CRM and service desk stop billing per seat, per conversation and per ticket, cloud egress disappears because no records leave the building, and idle seats no longer renew. The spend moves once into owned hardware instead of recurring every year into subscriptions.

Does citizen and case data leave the department?

No. Mickai runs fully offline on the department's own hardware. Prompts, records and results stay on site, and every action is sealed to the Open Audit Record, so there is no cross border transfer and no third party cloud holding public data.

Is Mickai certified to SOC 2 or ISO 27001?

Mickai produces the sealed evidence that supports SOC 2, ISO 27001 and GDPR examinations. It does not claim the certificate is held on your behalf; it gives your auditor a signed, on-premise trail to examine against those frameworks.

Which studios cover a public sector back office, and are they ready?

Pythia for analytics, Xenia for customer relationship management and Ponos for the IT service desk, all reached through the Omni Assistant. Mickai has sixty-three studios in total, ten production-ready at launch and fifty-three in development, with the back office set in the launch group.