Personal research project · not for commercial use

A prototype exploring how
Legionella risk assessment
could actually work.

A non-commercial proof-of-concept testing whether a dynamic, HSG274-grounded assessment workflow — aligned to how engineers actually think on site — could be built into a single tool. Developed in personal time, outside any working engagement.

Research & non-commercial use

Developed solely for technical research, process innovation, and skills development, the platform is not operated commercially due to professional conflict-of-interest considerations. It serves as a practical demonstration of my interest in applying emerging technologies, data analysis, and automation to solve complex compliance challenges. The project does not take clients, accept paid engagements, or solicit commercial relationships of any kind.

01What this project explores

Open questions the prototype is testing.

Working risk assessors typically use three or four separate tools to complete one job. The prototype tests whether a unified workflow — aligned to the engineer's mental model on site — is achievable as a single integrated tool, rather than a stack of integrations. None of these are claims; they are research questions.

Q1
Dynamic question generation

Can the assessor describe a system once and have only the relevant HSG274 questions surface? Today most tools render every question regardless of system attributes.

Q2
Engineer-aligned data model

Most platforms are built around assets and tasks because that is what the compliance manager wants to see. Would a system-mental-model data shape map closer to what the engineer is actually doing on site?

Q3
Single source of truth

Photos in one tool, temperatures in another, narrative in a third. Could all of it live in one data model from capture through to the final PDF, eliminating the assessor's data-integrator role?

Q4
Report as output, not document

Most reports are auto-generated from templates and feel generic. Could the report be a structured projection of the assessment itself, with engineer judgement interleaved into the structured output?

Q5
Action plan derivation

Findings of severity B/C should flow into the action plan automatically with HSG274-aligned timeframes. No double-entry. Testable as a small UX win or a substantive workflow change.

Q6
Sign-off as terminus

Sign-off is typically a parallel ceremony with a separate authoring pass. Could it be the natural final step of the on-site flow, with the PDF already done when the engineer leaves the site?

02Standards reference

Built against the documents the work actually cites.

Domain references the prototype works against. These are informational pointers, not claims of certification or accreditation.

01 / 04
ACOP L8 (4th Ed.)

The approved code of practice for Legionella control.

02 / 04
HSG274 Part 2 (2024)

Current technical guidance for hot and cold water systems.

03 / 04
LCA Standard 801.21

Service delivery standard for risk assessment.

04 / 04
BS 8580-1:2019

Risk assessments for Legionellosis — code of practice.

03Reading

Analysis and notes that inform the prototype.

Domain pieces written alongside the project — partly to document thinking, partly because the topic is interesting in its own right.

All blog posts →Reference guides →