custra
Platform overview

The full-building intelligence platform.

Detection across 24 systems. A complete service-case lifecycle. Eight plug-in capabilities. Expert escalation when the work requires it. Every piece composable — none of it actuating.

One case from fault to close

Every fault Custra detects opens a ServiceCase — a single governed object that tracks the work from first signal to verified closure. The case machine handles the bureaucracy so your team handles the building.

1

Detect

Atlas IS + OI identify the fault. Root-cause correlation folds coupled symptoms into one incident.

2

Triage

Criticality, regulatory flag, and spend estimate are set. Business-critical faults (leaks, fire, life-safety) auto-escalate.

3

Authorize

Policy-gated: auto / escalate-routine / escalate-critical. Your thresholds, not ours. A profile edit, not a code change.

4

Dispatch

Atlas MO sources the right trade by capability — HVAC to your HVAC contractor, leak to your plumber. Insource or outsource, per policy.

5

Repair

The tech is on-site. Custra’s copilot delivers the work guide, parts list, and safety tips — and escalates to a remote specialist when the job warrants it.

6

Verify

The system returns to baseline. Custra debounces the clearance signal before marking the fault resolved. Billing is gated on verification.

7

Bill

Verified work triggers the billing event. The CFO plug-in accumulates running dollar value; compliance log records the audit trail.

8

Close

Case closed. The event feeds the causal graph — every dispatch makes future detection smarter for this building and the whole fleet.

Escalation loop: If a critical case requires human authorization and the approver doesn’t respond, Custra cycles through the escalation chain — push, remind, hop to the next approver — until approved. Break-glass or hold-and-broadcast on exhaustion. The policy defines when; Custra enforces it without exception.

Eight plug-in capabilities

Every capability runs in the pipeline at the stage the surface map assigns. Add a new one by registering it — no core change required.

Fault diagnosis (FDD)

Root-cause enrichment at the DETECT stage. When a fault fires, FDD isolates the physical cause — a stuck damper, a failed bearing, a refrigerant shortfall — and names it in the advisory. Coupled live into the pipeline on every fault.

Technician copilot

At RECOMMEND, the copilot generates a work guide for the dispatched trade: the diagnosis, step-by-step instructions, tools, parts, safety flags, and senior tips from prior similar jobs. Delivered to the tech before they enter the plant room.

Risk ranker

Deferred maintenance scored at RECOMMEND: failure probability × replacement cost × regulated-system flag. The output is a ranked dollar list — the three items that most need attention, with a justification for each. Not a backlog; a decision.

Energy M&V

Verified savings tracking at VALIDATE. Baseline vs. actual energy consumption — the standard M&V protocol — attributed to specific repairs. The number you can put in a sustainability report and defend to an auditor.

CFO reporting

Running hard-dollar value at VALIDATE: verified repair cost vs. avoided damage, downtime cost vs. time-to-dispatch, and energy savings to date. Accumulates across the portfolio. The number the CFO actually wants to see.

Compliance log

Regulated-fault audit trail at VALIDATE. Every refrigerant event, fire-safety incident, and regulated-system fault is logged with the sensor reading, the advisory, the authorization, the dispatch, and the verification — ready for the inspector.

Digital twin

Asset model for the building: every system, component, and connection — typed to the Haystack and Brick schemas. The twin is the anchor for cross-building learning and the federated layer. Built to be federatable from day one.

Tenant management

Tenant intake and status at VALIDATE: SLA clock running from the moment the case opens, status updates to the tenant portal, and SLA-breach escalation if the clock runs out before verification. The tenant sees what you want them to see, when you want them to see it.

Specialist escalation — tier 2

When the copilot’s tier-1 guide isn’t enough — a critical fault, a regulated system, a diagnosis the on-site tech isn’t confident in — Custra escalates to a remote specialist on retainer.

The platform decides when escalation is warranted, routes to a capable expert by domain and language, and hands them the full diagnosis package on connect. The specialist advises; the on-site tech acts. Insight-only throughout.

The video bridge is an integration seam — Twilio, Zoom, or any compatible provider. No lock-in.

Customer-tunable policy

Every customer has a CustomerProfile that governs how Custra behaves for their buildings. Set it at onboarding; adjust it when your policy changes. A profile edit, not a code change.

Authorization tiers

Auto-approve below your spend threshold. Escalate-routine for scheduled repairs. Escalate-critical for life-safety and regulated systems. Break-glass thresholds for emergencies.

Escalation timing

Severity-scaled: a critical fault cycles faster than a routine repair. Remind intervals, hop logic, and exhaustion behavior are all profile-set defaults you override per customer.

Vendor thresholds

Confidence floor before dispatch, spend cap before escalation, preferred vendor list per trade, insource vs. outsource preference per capability class.

Regulated-system flags

Mark which systems in your building are regulated (refrigerant, fire, life-safety). Custra routes them through the full compliance path — log, escalation, audit trail — automatically.

Ready to see it in your building?