Anonymised delivery evidence

The capability matters more than the client logo.

This page is organised around the difficult software problems I have worked through: inherited workflow logic, regulated capture, operational case management, controlled application intake, large commerce estates and governed automation. Client identity and protected implementation detail are intentionally removed.

01
Enterprise workflowRecovery + production support

Recovering dense business rules in a long-running workflow system

Analysed a mature enterprise application where role, stage, region, task allocation, approval, inspection and invoicing rules had accumulated across application code, ORM behaviour and relational data. The work focused on reconstructing the actual operating model first, then applying narrow fixes without breaking continuity.

ProblemBusiness truth distributed across legacy code and data
ApproachRule reconstruction, trace-driven diagnosis and controlled refactoring
EvidenceProduction support, backlog reduction and repeatable recovery practices
02
Regulated onboardingCustomer journey

Turning a paper-style regulated application into a structured digital journey

Translated a detailed application process into a responsive customer flow with conditional capture, validation, consent-sensitive information, intermediary context and an operational submission boundary. The transferable capability is not the industry label; it is converting complex business rules into a journey that remains understandable to a user and operable by the business.

ProblemHigh-friction application with conditional requirements
ApproachProgressive capture, validation and consent-aware flow design
EvidenceCustomer-facing production delivery in a regulated environment
03
Case and matter operationsArchitecture specification

Designing a secure workspace around hundreds of operational matters

Authored a detailed technical specification for a matter-centred platform covering responsibilities, diary and reminder behaviour, documents, supervision, role boundaries, security, migration, testing, recovery and formal approval gates. The architecture starts with how work moves between people, not with a framework choice.

ProblemSpreadsheet and file-based operational administration
ApproachMatter-centred workflow, role oversight and document control
EvidenceReviewable multi-section technical specification with acceptance gates
04
Programme intakeControlled workflow

Open and invitation-only application workflows with configurable evidence rules

Delivered application-window behaviour, controlled invitation access, role-aware administration, programme-specific document requirements, validation and recovery support. The same pattern applies to grants, supplier onboarding, tenders, controlled registrations and other document-heavy intake processes.

ProblemDifferent programmes require different access and evidence rules
ApproachConfigurable windows, tokenised access and document contracts
EvidenceImplemented workflows, test evidence and operational recovery notes
05
Enterprise commerceOperations + modernisation

Operating and improving a large catalogue, quoting and procurement ecosystem

Worked across product discovery, large catalogue structures, collections, quote flows, account and supplier data, synchronisation, administration, accounting-connected processes, search architecture and incremental modernisation. The core challenge is keeping customer experience, internal operations and legacy data consistent while the platform evolves.

ProblemLarge product estate connected to operational workflows
ApproachData-aware commerce architecture and staged improvement
EvidenceTens of thousands of catalogue records and thousands of collections
06
Governed automationArchitecture + working patterns

Putting policy, approval and evidence around automated execution

Designed and implemented patterns where automation is given explicit purpose, bounded capabilities and clear authority. Read-only actions can proceed automatically, consequential changes require approval, destructive or disallowed actions fail closed, and release or execution claims are backed by evidence rather than inferred from source changes.

ProblemAutomation becomes risky when authority is implicit
ApproachPolicy states, human approval, exact release identity and audit evidence
EvidenceWorking control patterns, public demos and evidence-led delivery rules

Exercise the governed AI lab →

Transferable evidence

These are systems patterns, not industry souvenirs.

The same architectural skills reappear across different organisations: recover truth, model state, connect systems, protect boundaries, make operations visible and prove the release.

  • Business-rule recovery from code, data and runtime evidence
  • Progressive customer capture with validation and consent boundaries
  • Role-aware case, matter and approval workflows
  • Document-heavy intake with configurable requirements
  • Large operational data estates and staged modernisation
  • Human-governed automation with release and audit evidence
See the capability in motion

The demos are the next layer of evidence.

Use the labs to exercise synthetic versions of the patterns, then use the solution catalogue to map the pattern to a real engagement.