Predictive Maintenance Infrastructure

The decision layer for fleet operations.

Enterprise fleets already run telematics, maintenance systems, CMMS platforms, and supplier networks. Rooutiq sits between them — interpreting signals, determining what needs to happen next, and routing that decision into the systems already running the fleet.

Talk to RooutiqExplore the Platform
Systems of record
Remain yours
Deployment
Cloud, no hardware
Data sources
Existing telematics
Automation
Opt-in, policy-gated
Where Rooutiq sitssignal → decision → execution

  1. Telematics / diagnostics

    Signals the fleet already collects

  2. Rooutiq decision layer

    Fusion, vehicle context, scoring

  3. Prediction + evidence

    Failure mode, timing, confidence

  4. Work order / CMMS

    Routed into maintenance execution

  5. Procurement

    Parts, authorization, spend controls

  6. Fleet operations

    The vehicle back in service

The Category

Your fleet already has the data. It needs a decision layer.

Enterprise fleets already run telematics, diagnostics, maintenance systems, procurement and financial controls. Each one works. They just work as separate workflows, joined by people re-entering the same information. Rooutiq connects them.

  • Telematics
  • GPS
  • Diagnostics
  • CMMS
  • Work orders
  • Procurement
  • Supplier networks
  • Financial controls

Today, in most fleets

  1. 1Telemetry
  2. 2Alert
  3. 3Human investigates
  4. 4Work order
  5. 5Find part
  6. 6Get approval
  7. 7Purchase
  8. 8Repair

With Rooutiq

  1. 1Vehicle signal
  2. 2Emerging fault detected
  3. 3Evidence validates risk
  4. 4Maintenance action prioritized
  5. 5Correct repair path identified
  6. 6Procurement workflow initiated
  7. 7Spend controls enforced
  8. 8Maintenance completed

Rooutiq applies machine learning and multi-source signal fusion to determine which vehicle events require action. The output is not another alert stream — it is a prioritized decision, the evidence behind it, and the workflow that resolves it.

Reference architecture

Between the systems a fleet has and the decisions it needs.

Rooutiq is not a system of record. It reads from the fleet's systems and writes decisions back into them.

Rooutiq architecture: telematics and vehicle diagnostics feed the Rooutiq intelligence layer, which performs signal fusion, fault detection, risk scoring, prediction and evidence assembly. That produces a decision — what needs attention, why, when, which part and which action — which is executed through CMMS, maintenance, parts procurement and spend controls, producing operational outcomes.

Telematics

  • Geotab
  • Samsara
  • Verizon Connect
  • OBD / CSV
  • Vehicle diagnostics

Rooutiq intelligence

  • Signal fusion
  • Fault detection
  • Risk scoring
  • Prediction
  • Evidence

Decision

  • What needs attention
  • Why it matters
  • When to act
  • Which part
  • Which action

Execution

  • CMMS / work orders
  • Maintenance
  • PartsTech
  • Procurement
  • Spend controls

Outcome

  • Less downtime
  • Faster repair
  • Fewer repeat failures
  • Auditable decisions

Operating Model

From signal to execution.

A continuous operating loop, not a one-way alert pipeline.

  1. 01

    Sense

    Ingest signals from the fleet's existing systems — telemetry, diagnostics, fault codes.

  2. 02

    Understand

    Fuse telemetry, diagnostics, vehicle identity, maintenance history and contextual evidence.

  3. 03

    Predict

    Identify emerging failures and estimate operational risk, with stated confidence and evidence depth.

  4. 04

    Decide

    Determine what action should happen next — act, review, monitor, or blocked, with reasons named.

  5. 05

    Execute

    Route the decision into work orders, procurement, approvals and operational workflows.

  6. 06

    Control

    Govern every stage above by policy, authorization, evidence requirements, spend limits and audit.

Control is not the last step — it governs all five before it. Decision gates fail closed: where authorization, readiness or evidence is missing, no action is taken and the refusal is recorded with its reason.

The control model

Platform

One intelligence layer across the maintenance workflow.

Fleet status, vehicles needing attention, and open faults in one authoritative view — before anyone interprets a chart.

Decision record

Every decision leaves a record that can be audited.

An alert tells an operator something is wrong. A decision record states what was observed, what Rooutiq concluded, the evidence behind it, what it recommended, which controls applied, and what was actually done.

How decisions are governed

TRK-104DR-2214

2021 Freightliner Cascadia · VIN ending 4471

Decision recorded
What Rooutiq observed
  • NOx sensor circuit fault, provider-confirmed, 3 sightings
  • Exhaust temperature trend rising over 6 weeks
  • DPF regeneration service completed 7 weeks prior
What Rooutiq believes
  • Probable failure: NOx sensor (outlet)
  • Severity: high
  • Estimated window: 12 days

Confidence

81%

Reliability

Corroborated

Why Rooutiq believes it
  • 3 independent evidence groups fired
  • Telemetry trend + diagnostic fault + maintenance history
  • No conflicting evidence recorded
What Rooutiq recommends
  • Inspect within 2 weeks
  • Part identified: NOx sensor, outlet — fitment verified to VIN
  • Estimated repair cost: $640
What happens next
  • Work order raised in the fleet's CMMS
  • Parts request prepared, pending authorization
  • Assigned to the maintenance queue
What was actually done
  • Sensor replaced, 9 days after the decision
  • Outcome verified by the technician
  • Prediction resolved against the repair record
What controls applied
  • Automated procurement: not enabled for this fleet
  • Spend authority: not verified — purchase left to a human
  • Policy version and readiness state frozen on the record
What was recorded
  • Immutable decision record with its full input snapshot
  • Actor, model version and policy version stamped
  • Retained for audit; never rewritten by a later decision

Illustrative record. The values above demonstrate the structure of a Rooutiq decision record. They are an example, not a customer outcome or a published accuracy figure. Confidence is a model output on its own scale — empirical accuracy requires verified repair outcomes and is reported separately.

Ecosystem

Built to fit into the systems fleets already use.

Rooutiq sits between the data a fleet already produces and the systems it already runs — no hardware, no replacement of the maintenance stack.

TelematicsRooutiqMaintenancePartsProcurementCompletion

Telematics

Geotab · Samsara · Verizon Connect · OBD / CSV

CMMS / Work Orders

eWorkOrders

Parts & Procurement

PartsTech

Enterprise Systems

Fleet REST API · Google Sheets export · Stripe billing

Available today

  • Parts identification from the predicted failure mode
  • Live supplier availability and pricing
  • Parts ordering against a prediction
  • Work orders pushed to the CMMS, status read back
  • Repair outcomes verified against the original prediction

Roadmap

  • Coordinated shop scheduling
  • Automated billing reconciliation
See supported integrations Named systems are data sources Rooutiq can read from. They do not imply certification or partnership.

Enterprise Control

Intelligence is only useful when enterprises can control what happens next.

Automated procurement is not “the model predicted a part, so money moved.” Prediction, decision, authorization and execution are deliberately separate concerns, and a purchase has to satisfy every one of them.

01Prediction
What the evidence says about a vehicle.
02Decision
What should happen, and how confident that is.
03Authorization
Whether this organization may act on it at all.
04Execution
The action itself, recorded with its basis.

Every automated purchase traverses this chain

  1. 01Prediction confidence
  2. 02Evidence requirements
  3. 03Procurement policy
  4. 04External account verification
  5. 05Spend authority
  6. 06Approval
  7. 07Execution
  8. 08Audit trail

A failure at any step stops the purchase and is recorded with the reason. Nothing proceeds on a step that could not be evaluated.

Fail-closed by default

When a required condition is unresolved, no purchase occurs. Absence of evidence is never treated as permission, and automated purchasing is off unless explicitly configured.

Server-side authorization

Workspace, actor, readiness and eligibility are resolved server-side. Values asserted by a browser are discarded rather than merged.

Tenant isolation

Operational records are workspace-scoped, and cross-workspace writes are refused rather than filtered after the fact.

Auditable decisions

Each decision retains the actor, workspace, verdict, blocking reasons and the rule version in force at the time, so a refusal from last quarter is not read under today's rules.

Verified external relationships

A connected financial or supplier account is not authority to spend. Purchasing authority is verified independently before it counts.

Separate spend ceilings

Per-transaction and aggregate limits are enforced independently, so a systematic signal error cannot pass by staying under a single-ticket cap.

Rooutiq does not hold SOC 2, ISO 27001, HIPAA or FedRAMP certification. These are architectural properties of the platform, stated as such — not compliance claims.

Enterprise Scale

One intelligence layer across the fleet.

The system is designed to operate at organizational scale — not as a separate workflow per vehicle, per site, or per integration.

  • Organization-level data isolation
  • Fleet-level automation policy
  • Role-based access and directory sync
  • Centralized visibility across sites
  • Auditable decision records
  • Controlled, opt-in automation

  1. 100 vehicles

  2. 1,000

  3. 10,000

  4. 50,000+

What each team gets.

Fleet operations

One prioritized queue of vehicles that need action, ranked by risk and evidence instead of alert volume.

Maintenance

The likely failure mode, the supporting evidence and the repair path — delivered into the work-order system the shop already uses.

Finance

Per-transaction and aggregate spend ceilings, explicit authorization, and a kill switch that stops automated purchasing without stopping detection.

IT

Reads from existing telematics and maintenance systems, exposes a REST API and SCIM directory sync, and adds no hardware.

Procurement

Decisions resolved to the specific vehicle and part, routed into the supplier workflow with the purchase attributed back to the prediction.

Risk and safety

Every decision retains its actor, evidence, verdict and rule version, so an action taken months ago remains explainable.

See how Rooutiq fits into your fleet architecture.

Connect your existing telematics and operational systems, evaluate the decision layer, and see how predictive maintenance moves from detection to execution.

Talk to RooutiqExplore the Platform