MapFlow — LinkedIn Post Series

UK councils: if your planning, social care or finance records sit inside a US-parent SaaS platform — even one with a "UK region" and a watertight DPIA — you are exposed to the US CLOUD Act.

The CLOUD Act lets US law enforcement compel American cloud providers to hand over data regardless of where it's stored. As Kiteworks put it publicly:

"Organisations using US cloud providers—regardless of 'UK regions' or contractual data protection promises—surrender jurisdictional control to American corporations subject to US legal authority."

Contractual clauses do NOT resolve this. US statutory obligations override your contract. "UK regions" do not resolve it. The provider answers to US law first.

The only architectural fix is genuine sovereignty:
✅ Customer-managed encryption keys held outside the provider — compelled disclosure yields only ciphertext
✅ UK sovereign deployment with no US corporate entity in the relationship

That is exactly what MapFlow ships. Citizen data is processed on a council-owned local estate (a Mac mini host or on-prem proxy) running sovereign LLMs — Llama 3, Qwen, Mistral — with AES-256-GCM keys bound to the device Secure Enclave. No US-parent platform ever sees plaintext citizen data. The CLOUD Act cannot reach a relationship in which no US entity exists.

Sovereignty is a property of architecture, not of contract wording.

MapFlow is the only UK sovereign data-migration platform that combines AI field mapping with zero CLOUD Act exposure.

Sources: Kiteworks, "The CLOUD Act and UK Data Protection: Why Jurisdiction Matters"

#CloudAct #DataSovereignty #LocalGovernment #UKGov #InfoSec #MapFlow

The LGA's 2025 "State of the sector: AI" survey of English councils tells the story in numbers:

• 95% of councils are using or exploring AI
• Realised benefits: staff productivity (36%), service efficiencies (33%), cost savings (21%)
• But the three biggest barriers are: lack of funding (62%), lack of STAFF CAPABILITIES (56%), lack of STAFF CAPACITY (52%)
• 66% can't even scope where AI would add value

The binding constraint is not appetite. It is capacity. Councils have lost the in-house technical staff to specify, scope and deliver.

The Tony Blair Institute: "AI could release staff from repetitive tasks and relieve an overburdened and demotivated workforce."
The Public Accounts Committee: "AI has the potential to radically change public services by automating routine tasks."

Data-field mapping IS that repetitive task. Mapping a legacy Idox/Uniface field to a Salesforce/Arcus or Microsoft Dynamics field, classifying a column's PII category, generating a table purpose statement — these are exactly the jobs that burn out the few technical staff councils have left.

MapFlow's AI is built for this. Fine-tuned LoRA adapters trained on de-identified structural metadata (field names, types, labels — never raw citizen values) learn the council's OWN estate. Every mapping an officer accepts, every PII call they confirm, every screenshot correction becomes a training pair. The next migration is more accurate and less manual.

The AI is explicitly allowed to learn — on the council's own sovereign estate — so the technical staff the LGA identifies as the bottleneck are unburdened, not replaced.

Let AI carry the load the sector can no longer staff.

Sources: LGA State of the sector: AI 2025 update; Tony Blair Institute; UK Parliament Public Accounts Committee.

#ArtificialIntelligence #LocalGovernment #DigitalTransformation #UKGov #MapFlow

For years, the only politically viable answer to a council leaving a 30+ year-old client-server system was a bare-bones "lift and shift" onto the cheapest SaaS bid. The public record now shows exactly what that produces.

HILLINGDON — EY external audit (Audit Committee, 10 Feb 2026):
• "Issues with the implementation of Oracle Fusion and Oracle EPM were noted, leading to inaccurate user lists, incomplete data, and a reliance on manual interventions by the finance team."
• "Gaps were identified in the Council's arrangements for system implementation, including a lack of clear 'go-live' criteria, insufficient user acceptance testing, and inadequate monitoring of privileged user access."
• Result: a DISCLAIMED audit opinion — the most serious finding a local authority auditor can issue.

HAVANT & BRACKNELL — planning system "lift and shift" onto Arcus Global:
The Register (6 May 2026): "Three UK councils have hit problems processing planning applications and land searches following a systems migration to a new software-as-a-service provider." House sales fell through. A 5G mast went up in a conservation-area cricket ground. Hundreds of stalled land searches.

Havant Civic Society: "In April 2024, Havant Borough Council had knowingly signed a contract for a bare-bones solution with no data migration and no local customisation. Bracknell Forest Borough Council, it appears, may have fallen into the same trap."

"The database itself – the planning history – was missing completely at launch because the essential data migration, conversion and load had not been carried out prior to the production cutover."

The diagnosis that matters: "the councils' weak specification and limited understanding of the full data and system migration made it far more likely that they mis-bought than that Arcus mis-sold."

Councils MIS-BOUGHT. Because they no longer hold the in-house capacity to write a requirements spec or run a data migration. The political pressure to "just go live" produces the worst engineering the politics will allow.

Lift and shift is over. The migration must be engineered first.

Sources: Hillingdon Audit Committee 10 Feb 2026 (EY); The Register 6 May 2026; Havant Civic Society 22 April 2026.

#LiftAndShift #LocalGovernment #PublicSectorIT #Audit #Hillingdon #Havant #Bracknell #MapFlow

Pull the third-party evidence together and one need emerges that no incumbent is meeting:

1️⃣ CLOUD Act: US-hosted SaaS is a sovereignty risk councils cannot contractually resolve (Kiteworks)
2️⃣ Capacity: 95% of councils want AI, but 52% lack the staff capacity to deploy it (LGA 2025)
3️⃣ Lift & shift: produces disclaimed audit opinions and missing planning histories (EY / Hillingdon; The Register / Havant & Bracknell)
4️⃣ Political mandate: AI "could release staff from repetitive tasks and relieve an overburdened and demotivated workforce" (Tony Blair Institute)

Every one of these is a barrier a different vendor has hit and recoiled from — back to the politically safe, technically worst option. Havant, Bracknell and Hillingdon are what that worst case looks like in production.

MapFlow is built to satisfy all four at once:
🔒 Sovereign local-estate AI — no US entity in the relationship, keys bound to the Secure Enclave
🧠 AI that learns the council's own schemas on its own estate, unburdening technical staff
⚙️ Migration engineered BEFORE it is political — data cleansed and mapped before cutover, not after
📊 A CLOUD Act posture that survives audit

This is why partners can sell MapFlow as the answer to a requirement everyone else has failed to meet — not because the need was unclear, but because sovereignty, capacity and lift-and-shift inertia made the worst-case solution the only viable one for years.

MapFlow is the best engineering the politics will finally allow — the only UK sovereign data-migration platform that combines AI field mapping with zero CLOUD Act exposure.

#LocalGovernment #DigitalTransformation #SovereignAI #PublicSector #MapFlow #Salesforce #Migration

For years, Havant, Bracknell and Hillingdon have been the only "viable" answers a council could politically defend: cheapest tender, no data migration, go-live and hope.

That is the worst engineering. It is what the politics used to allow.

MapFlow changes what the politics allows — by making the best engineering the safe choice:

✅ Sovereign — CLOUD Act cannot reach a relationship with no US entity
✅ Learning — AI carries the mapping load the sector can no longer staff
✅ Engineered — data cleansed and mapped before cutover, not after
✅ Auditable — a posture that survives EY-level scrutiny

If you're a council leaving a 30+ year-old client-server estate, a partner selling into local government, or an officer who's watched a lift-and-shift fail — let's talk.

Read the full white paper: the best engineering the politics will allow.

#MapFlow #LocalGovernment #PublicSectorIT #SovereignAI #DigitalTransformation

I've spent my career moving public-sector and corporate data. And the strangest thing has happened: the industry went backwards.

THE ENTERPRISE ERA — we did this properly.
Early in my career I ran Streetworks migrations. The moment Microsoft Access hit the relational complexity and transaction volumes, the Jet/ACE engine folded — the GUI tool had to be abandoned mid-project and we dropped to raw SQL scripting just to keep the migration alive. That was the warning shot.

Then National Highways: roughly one-third of the entire National Highways Assets register — high-volume geospatial and asset metadata — was mapped, validated and transferred on Sagent Data Flow. Not a single corruption event. That is what "enterprise ETL" meant: high-volume handling, strict schema enforcement, full data lineage.

THE COMPLIANCE BENCHMARK — corporate banking.
In highly regulated banking environments where data lineage and security are legally mandated, the industry standard relied on heavy-duty integration engines like Talend. Security, auditing, and multi-user concurrency were treated as foundational requirements, not optional luxuries. This was the world of KYC — Know Your Customer — where every field movement, every transformation, every access had to be lineage-traceable and audit-defensible. That was the benchmark we were all measured against.

THE REGRESSION — and then we went backwards.
Somewhere along the way, procurement frameworks decided that modern, cloud-native ETL SaaS was "too risky" — CLOUD Act panic, certification paywalls, certification technicalities — while a 1992-era desktop file database was "grandfathered" into the corporate desktop image and signed off without a single question.

So data specialists who stream 4K video on their phones and draft emails with AI are handed Microsoft Access to migrate millions of sensitive citizen records. A tool older than the smartphone, the smart speaker and the modern web. No upgrade path. No concept of a REST API. Structurally incompatible with modern cloud compliance. The "poor man of IT."

That is not progress. That is regression engineered by the very frameworks meant to keep data safe. The vendors left holding the bag — doing the best possible job with the tool they were forced to use — get blamed for outcomes the framework imposed on them.

THE SLEDGEHAMMER — I built the way out.
I built MapFlow to break the deadlock. The power of an enterprise ETL pipeline — lineage, governance, scale — inside a secure, sovereign, browser-isolated package. Real PostgreSQL 16 compiled to WASM. Zero-Trust by default. TPM 2.0 key binding. AES-256-GCM field encryption. Automated SIEM auditing. Sovereign AI that learns the council's own estate. Zero CLOUD Act exposure, end to end.

The enterprise era did it right. The regression was forced on us by gatekeeping, not by engineers. MapFlow is the sledgehammer that breaks the deadlock — and restores the enterprise benchmark on a modern, sovereign PaaS.

If you've watched a lift-and-shift fail, or been forced to ship an .accdb when you knew it was wrong — let's talk.

#DataMigration #ETL #Sagent #Talend #NationalHighways #MicrosoftAccess #Regression #LocalGovernment #PublicSectorIT #SovereignAI #MapFlow