One platform, from raw sensor data to a fixed building

Veil ingests messy BMS/HVAC data, tags it semantically, and runs your analytics rules continuously — surfacing faults, waste, and diagnostics to the people who fix them.

How a fault gets caught — and fixed

Every feature below plugs into this one pipeline, running continuously.

1

Connect

A BACnet connector, CSV upload, or HTTP push brings sensor readings in from the building automation system.

2

Tag

AI-assisted Haystack tagging (human-approved, always) labels what each sensor is, so rules find equipment by type, never by point name.

3

Evaluate

The Rules Studio runs every deployed rule against live and historic data, continuously, across every matching piece of equipment.

4

Alert

Spark opens an alert with full evidence, a technician works it through verification and close, with a threaded team comment trail.

5

Report

Every fault's cost is calculated and frozen at the moment it fired — auditable in a PDF report, and tracked fleet-wide in the KPI view.

The core: Spark & the Rules Studio

Everything else in Veil supports this one idea.

Spark — alerts & technician workflow

Every fault Veil finds, with a full lifecycle from first detection through verification and close, automatic post-fix verification, and a threaded team comment trail. Three views — calendar timeline, list, and table — all filterable and exportable.

Rules Studio

Write a rule once — in plain English, or by hand for full control — and it evaluates against every matching piece of equipment, across every building, forever. Rules find equipment by what it is (Haystack tags), never by point name, so one rule genuinely covers ten thousand buildings. Backtest against 90 days of history before you ever turn a rule on live.

Beyond fault detection

KPI

Ongoing performance tracking, distinct from fault alerts — energy per square meter, average power, runtime hours — computed daily per building and compared fleet-wide in a single pivot view.

Connectors

Automatically pull live sensor readings from a real building automation system. BACnet/IP is the first connector — point Veil at a device you already know the IP of, and a small agent handles the rest over an ordinary outbound connection. No inbound firewall changes, ever.

Explorer & Analysis

Find any sensor by what it is, not its exact name, and plot it instantly. Analysis (a free-form workspace, internally "Compare Canvas") overlays any combination of sensors from any building, with rule-firing windows shaded directly on the chart.

Reports & ROI

PDF reports in English or French — alert summary, top faults, mean time to resolve, estimated cost of everything still open, an equipment heatmap. Every alert's cost is calculated from a real formula and frozen at the moment it fired, so every number in every report stays auditable.

AI tagging, advisory only

Claude-assisted tag suggestions and rule drafting speed up setup — but nothing an AI suggests is ever applied automatically. Every suggestion waits for a human to approve it, and every change is logged in a full audit trail.

Team & white-label

Role-based access for consultants, technicians, and managers, per-person language (English/French), and white-label branding — your logo and colors, applied live across the app and baked into every PDF report.

Ready to see it on your own data?

Request a demo

Tell us about your portfolio and preferred deployment mode — we'll set up a walkthrough.

Sent directly to our team — we typically reply within a business day.