AI-enabled inspection platform for a regulated industry

Context

Building inspection in Québec is a regulated profession working under the RBQ framework. DomiSpect is a SaaS platform built and operated by my team for inspectors who must produce structured, traceable reports after each visit — often on sites with no connectivity.

Challenge

Inspectors were spending four to six hours after every visit producing a report aligned with the applicable inspection framework and sorting hundreds of photos and pieces of evidence by hand, while their professional liability required every finding to be documented and traceable.

My role

Founder: product vision, architecture and technology leadership, from the first prototype to the live platform.

Architecture / approach

Build the report during the inspection itself rather than after it: structured dictation mapped to the regulatory report structure, photo and evidence management attached to each finding, an AI review pass before signature that flags gaps and inconsistencies, and a fully offline mode that synchronizes when the connection returns.

Technology

Cloud SaaS with a mobile application; AI-assisted report generation and review; document and photo pipelines; offline-first synchronization; bilingual product designed to support regulated inspection processes.

Execution

Delivered end to end by a small, focused team — from prototype to live platform: report structure, evidence management, AI review pass and offline mode.

Constraints

Sites without connectivity; a report structure imposed by the inspection framework; every finding must remain traceable to its evidence; no claim of legal compliance — the product is designed to support the regulated workflow, not to certify it.

Key takeaways

1. Capture structure at the source — during the inspection — and the report writes itself.
2. An AI review pass is most valuable right before signature, where the cost of an omission is highest.
3. Offline-first is a field requirement, not an optimization.