The AI your operations actually run on.
Predictive maintenance, anomaly detection, and autonomous triage across solar, wind, storage, and distributed industrial assets. Built around your SCADA, your historian, and the technicians you actually dispatch.
PalantirVanguard Elite Partner
GeminiGemini Enterprise AI Practice
OpenAI
Anthropic
AWS
Microsoft
Databricks
Snowflake- 14:31WND-B-247 · curtailed to 60%, WO #4471 routed
- 14:28BESS-D-04 · cycle complete, returning to standby
- 14:14SOL-A-118 · soiling flagged, WO #4470 drafted
- 14:02INV-C-12 · thermal recovered
Our work with Sonnedix
Asset intelligence, deployed across one of the world's largest independent solar IPPs.
“Northslope is a partner, but is really embedded in our team. And that’s really the most important thing for us.”
“We are now getting to the point where every day I can see what has been going on. I can track the progress without having to do the work, because the data is flowing through automatically.”
SCADA / Historian
OSIsoft PI, AVEVA, GE Proficy
Vibration & Acoustics
Bearings, gearboxes, pumps
Inverter Telemetry
String, MPPT, thermal
EAM / Work Orders
Maximo, SAP PM, Fiix
Weather & Grid
Forecast, curtailment, prices
Manuals & History
OEM specs, prior failures
30-day risk forecast
thermal stress
bearing degradation
preventive work orders
revenue at risk
matches 3 prior fleet events
Anticipated failures
14–28 days before symptom onset.
Work orders drafted
Routed to the right tech with parts and procedure.
Underperformance flagged
Soiling, shading, miscalibration spotted in cohort drift.
Pattern library
Every failure becomes a fingerprint for the next one.
We're AI engineers who sit in your ops centre.
You know your assets. We know how to make AI work alongside the people watching them. Our Forward Deployed Engineers embed with your reliability and tech-ops teams, understand the failure modes you actually see, and ship agents that earn trust before they earn autonomy. Built around your fleet. Supported forever — consumed like SaaS.
Built for the reliability engineer at 2am, the dispatcher at noon, and the VP at the QBR. Every screen earns its pixel.
Fleet Command
Geographic health, region output, revenue at risk, and the live event stream — all on one surface. Cluster, drill, dispatch. Drift surfaces as colour, not as a buried alert email.
Asset Deep-Dive
Production, revenue waterfall, meter hierarchy and peer-ranked performance — every asset evaluated against its cohort. The view a reliability engineer reaches for first.
Agent-Drafted Work Order
Schedule, parts staged with stock + cost, technician assigned, scope checklist, and dispatch actions — the agent drafts, the human approves. One surface, no ticket assembly across four systems.
Parts Intelligence
The truck never rolls without the part. Live ERP sync, reorder thresholds, supplier and class — wired directly into the work-order draft so dispatch knows what's staged before it leaves the DC.
Six core capabilities, composed for your asset class. Built around your historian, your EAM, and the technicians you actually have on the ground.
Predictive Maintenance
Failure-pattern models trained on your historian and OEM data — catching degradation 14–28 days before traditional alarms.
Custom to your asset class- InputsHistorian tags, CMMS history, OEM bulletins
- ModelsPer-class survival + cohort drift
- OutputProbability-of-failure ranked by revenue at risk
Anomaly Detection
Cohort-based outlier detection across your fleet — flagging the inverter that's drifting before it crosses any threshold.
Custom to your fleet shape- InputsStreaming SCADA, sub-string telemetry, weather
- MethodCohort residuals, not static thresholds
- LatencySub-15-minute on streaming data
Autonomous Triage
Agents draft work orders with diagnostics, parts and procedure attached. Humans approve, not assemble.
Custom to your EAM- Writes toMaximo, SAP PM, Fiix, IFS
- DraftsPriority, fault code, parts BOM, tech match
- AuditEvery action logged with reasoning trace
Performance Analytics
String-by-string, turbine-by-turbine attribution — telling you why availability dropped and where to recover it.
Custom to your KPIs- InputsString / turbine telemetry, irradiance, wind
- MethodLoss-tree attribution by failure mode
- OutputPer-asset PR + recoverable MWh quantified
Technician Dispatch
Route optimisation across truck rolls, parts inventory, and certification. The right human, parts, the first time.
Custom to your geography- InputsRoster, certs, vehicles, inventory, SLAs
- OptimisesTruck rolls, cert match, batched routes
- DeliveryPushes to your existing field tech app
Reliability Library
Every failure, root cause, and fix becomes searchable institutional memory. Your senior engineers' knowledge, retained.
Custom to your fleet history- InputsClosed WOs, RCAs, OEM bulletins, tech notes
- IndexingVector + structured, signal-pattern fingerprints
- QueryNatural language, ranked and cited results
Generic predictive maintenance vendors model "machines." We model the assets you operate, the way they degrade, and the playbooks your team already trusts.
Solar (utility-scale)
Inverter strings, trackers, soiling, and curtailment events compounding across thousands of assets. We model cohort drift across the array, separate weather noise from real degradation, and feed wash schedules and tracker recalibration directly into your O&M plan.
String Underperformance
Detect drift below cohort 8% earlier than threshold alarms.
Soiling & Wash Optimisation
Schedule washes against weather and price forecasts.
Tracker Diagnostics
Identify miscalibrated rows from cross-row variance.
Wind
Gearbox bearings, blade leading edge, pitch systems, and curtailment trade-offs. We integrate vibration analysis with SCADA and met-tower data to predict failures with weeks of lead time, when you can still get a crane on a calm-weather window.
Gearbox Wear Detection
14–21 day warning on bearing degradation.
Blade Health
Edge erosion and ice events from acoustic + power signals.
Curtailment Optimisation
Decide when to derate vs. take the wear hit.
Energy Storage
Cell-level health, thermal runaway risk, cycling strategy and warranty management. We integrate BMS telemetry with market signals to optimise revenue without compounding degradation.
SoH Forecasting
Per-rack state-of-health vs. warranty curves.
Thermal Anomaly Detection
Pre-cursor patterns to runaway events.
Cycle Strategy
Trade revenue vs. degradation per market interval.
Distributed Industrial
Pumps, compressors, motors, and chillers across hundreds of sites where no two assets are configured the same. We unify telemetry across vendors and build per-site failure models that respect the physics of the install.
Pump & Motor Health
Vibration + current signature analysis.
Multi-Site Dispatch
Optimise truck rolls across geography & parts.
Vendor-Agnostic Telemetry
Unify mixed OEM fleets in a single model.
Hydro & Run-of-river
Older fleets, sparse instrumentation, and decades of tribal knowledge that's about to retire. We build the institutional-memory layer alongside the predictive models.
Sparse-Data Modelling
Make use of low-frequency and analog signals.
Knowledge Capture
Encode senior operators' playbooks before they retire.
Outage Planning
Coordinate maintenance with hydrology forecasts.
Grid & Substation
Transformers, switchgear, and protection relays where one failure becomes a regulator's phone call. We integrate DGA, partial-discharge, and SCADA into per-asset risk scores tied to your asset-management plan.
Transformer Health
DGA + load-cycle modelling for end-of-life forecasts.
Partial-Discharge Triage
Separate noise from precursor events.
Asset Risk Scoring
Plug into regulatory asset-management frameworks.
Forward Deployed Engineers embed with your reliability and tech-ops teams, learn your failure modes, and ship agents that earn trust before they earn autonomy.
Sit in the Ops Centre
We work alongside your reliability engineers and tech-ops leads. We learn the failure modes you've actually seen — not the ones in the OEM manual.
Connect the Signals
SCADA, historian, EAM, weather, market data. We adapt to your stack — OSIsoft PI, AVEVA, Maximo, SAP PM, custom. Whatever's there.
Earn Autonomy
Agents start as advisors, not actors. Once they've proven themselves on your fleet, we extend their authority — case by case, with a kill switch.
See it with your fleet.
Your assets, your historian, your team. We'll connect to a slice of your fleet and show you what an agent finds in the first week.
Book a Demo