Applied AI · Mechanical reliability · Evidence-linked reasoning
An AI-assisted workspace that helps engineers organise equipment evidence, compare likely causes, and choose the next physical test.
A pump, motor, or bearing problem rarely arrives as a clean question. Measurements, maintenance notes, manuals, and competing explanations need to be considered together. A plausible answer is not enough: the engineer needs to see its evidence, what remains uncertain, and which observation would distinguish the remaining causes.
We built a workspace that keeps the equipment, observations, possible causes, and next test together as an investigation develops. It combines document retrieval, mechanical calculations, and AI-assisted reasoning, with an equipment view that links explanations to their evidence.
The deployed application brings intake, investigation, calculations, and evidence review into one workflow. Engineers can examine the reasoning, continue the investigation as observations arrive, and record physical findings.
Engineers retain control of the investigation. Recommendations link to their sources, and root-cause confirmation requires physical evidence.Executed software checks, source revision, reproduction commands, and evaluation scope.
An access token is required to run investigations; the public entry screen is available without one.
Public presentation of the pump-diagnosis workflow, August 2026.
Measurements, source passages, and calculations retain distinct provenance. The workspace preserves evidence snapshots and tool inputs so an engineer can inspect what a result depends on instead of treating generated prose as authority.
Versioned tools cover rotating-equipment calculations such as shaft and bearing frequencies, gear mesh, pump affinity, NPSH margin, thermal growth, and first-pass signal analysis. Their assumptions and inputs remain part of the investigation record.
Competing mechanisms stay open when evidence is insufficient. The next observation is chosen to help distinguish them, and a confirmed root cause requires physical inspection or a comparable corrective intervention. The equipment view connects diagnostic reasoning to the relevant machine components.
A FastAPI endpoint loads the investigation and hashes attachments. Retrieval selects published, tenant-permitted evidence; registered tools calculate explicit numeric inputs. The model proposes a structured next turn, the server validates it and enforces the confirmation boundary, and PostgreSQL retains the result with its evidence and tool records. A deterministic mapper produces the equipment view without accepting model-authored geometry.
On September 8, 2026, a clean source snapshot passed 47 automated tests and 16 known-answer calculation checks. Five synthetic, model-off continuity scenarios also passed their development rubric, including a case with insufficient equipment information. These checks cover software behavior, arithmetic, and evidence-only fallback handling. The dated verification record documents the source revision, test conditions, measured outputs, and evaluation scope.
When model synthesis is unavailable, the workspace continues in a clearly labelled evidence-only mode. Equipment-specific conclusions require sufficient equipment identity, and confirmed diagnoses require physical evidence. These controls keep the investigation traceable as new observations arrive.