Our work

Selected systems and applied work.

We label each project by its real stage so a working prototype is not mistaken for a live client platform.

Soft launch / demonstration

Asset survey and data-enrichment platform

Problem: Survey photographs, labels and basic records remained incomplete and hard to reuse after the immediate project.
System: A mobile-first survey and searchable asset record. AI uses photographs, labels, specifications and documents to enrich records and prepare decision information for review.
What it demonstrates: Image and document interpretation, structured enrichment, reporting and the separation of field capture from heavier AI processing.

Live managed service

Emergency-lighting compliance workflow

Problem: Recurring test data, faults, remedial actions and certification needed to become one continuous compliance record.
System: Scheduled extraction and reporting, fault identification, remediation logging, annual certification and technical escalation.
What it demonstrates: Operational compliance automation, recurring reporting, exception surfacing and evidence continuity.

Working internal prototype / in development

Compliance evidence engine

Problem: Compliance evidence is distributed across storage systems and difficult to monitor consistently.
System: The prototype ingests selected SharePoint and local folders and maintains an audit log. The wider design adds classification, evidence links, gap and freshness monitoring, alerts, reporting and access control.
What it demonstrates: Evidence-system architecture, a disciplined AI/rules split and auditability as a core requirement.

Internal pilot

AI-assisted sales agent

Problem: Prospect research and relevant outreach required significant founder time and careful control of sending quality.
System: An internal agent identifies prospects, prepares evidence-based drafts and tracks responses. Every message is reviewed and sending volume is deliberately constrained.
What it demonstrates: Research agents, evidence-grounded drafting, human approval and rate limits.

How we present a project

Every project write-up follows the same structure: stage tag, challenge, users, information sources, workflow, AI tasks, deterministic controls, human review, current result, limitations and next step. Client names, sites, volumes and screenshots appear only with written permission and redaction review.

Recognise one of these problems in your own operation?

Bring us the process. We will tell you where AI is useful, what should remain rule-based and what a practical first build would look like.

Discuss a workflow

No transformation programme required. Start with one process.