Case Studies / Elevator Manufacturer
MANUFACTURING & IoT · PREDICTIVE MAINTENANCE
Replacing manual elevator inspections with AI-driven sensor analytics
An elevator manufacturer needed to understand the real-time status and position of elevators in the field – data that was previously only available through manual inspections or service calls. Working with an IoT partner, we built ML models that extract location and state information from minimal sensor data, laying the foundation for predictive maintenance, even in environments where connectivity is unreliable.
Client: Elevator manufacturer (under NDA) – delivered in partnership with an IoT hardware company.
KEY RESULTS
Minimal
Designed for fewest possible sensors – scalable across fleet
Offline
Works in connectivity-limited environments (Faraday cage)
ML
Clustering + custom logic for state and position detection
PdM
Foundation for predictive maintenance at scale
INDUSTRY
Manufacturing
USE CASE
Elevator location & state detection
AI APPROACH
Clustering + custom logic
DATA
IoT sensor data (minimal sensor set)
CONSTRAINT
Limited connectivity / offline capable
GOAL
Predictive maintenance foundation
The challenge
Elevator maintenance has traditionally been schedule-based – technicians inspect units at fixed intervals regardless of actual condition. The manufacturer wanted to shift toward predictive maintenance: using real-time data to understand what each elevator is doing, detect anomalies early, and dispatch service only when needed. This would mean fewer unnecessary service visits, less downtime, and lower operational costs across large building portfolios.
The first step was a hard technical problem: reliably determining the position and operational state of an elevator using sensor data. And the constraints were significant. Elevator shafts create a Faraday cage effect that disrupts wireless connectivity. Cloud connections drop or delay unpredictably. The solution needed to work with minimal sensors to remain economically scalable across thousands of units. And there was no labeled ground truth data to train on – the system had to learn patterns from raw sensor readings.
The core problem: the manufacturer had no automated way to know what their elevators were doing in the field. Manual inspections were the only source of status information – expensive, infrequent, and reactive. The physical environment made even basic connectivity a challenge.
What we built
Working alongside an IoT partner who handled the hardware and device design, we focused on the data and intelligence layer – turning raw sensor signals into meaningful operational information.
Data collection protocols and preprocessing. We designed the data collection approach from scratch – defining what to capture, how to structure it, and how to preprocess noisy sensor data into usable datasets. This included handling gaps and delays caused by the intermittent connectivity in elevator environments.
ML-based position and state detection. Using advanced clustering algorithms enhanced with custom domain logic, we built models that determine the elevator’s current floor position and operational state – moving up, moving down, idle, door open, door closed – from minimal sensor inputs. The algorithms were designed to work without ground truth labels, learning patterns directly from the data.
Offline-capable architecture. The solution architecture was designed to function when cloud connectivity is unavailable – processing and storing data locally, then syncing when the connection is restored. This was essential for real-world deployment where elevator shafts routinely block wireless signals.
Minimal sensor design for scalability. The entire approach was optimized for the fewest possible sensors per elevator. Adding extensive instrumentation to every unit in a large fleet isn’t economical. We proved that meaningful operational intelligence could be extracted from a deliberately constrained sensor set, making fleet-wide deployment viable.
The results
BEFORE
No automated visibility into elevator status. Maintenance was schedule-based. Service calls were reactive. No data infrastructure in place.
AFTER
Automated position and state detection from sensor data. Offline-capable architecture. Minimal sensor footprint designed for fleet scalability. Foundation for predictive maintenance.
The project delivered the data collection, processing, and ML infrastructure needed to understand elevator behavior in the field – the foundational layer for predictive maintenance. The manufacturer gained automated visibility into elevator operations that previously only existed through manual inspection.
The technical approach – extracting meaningful intelligence from minimal sensors in connectivity-constrained environments without labeled training data – demonstrated theBlue.ai’s ability to work within real-world industrial constraints, not just clean lab conditions.
Technology used
More Case Studies
See how we’ve helped other companies
AUTOMOTIVE (UNDER NDA)
Replacing reactive maintenance with a predictive maintenance roadmap
Unplanned equipment failures were causing costly production stops. We assessed the manufacturer’s infrastructure and delivered a concrete architecture for predictive maintenance, a roadmap from reactive repairs to data-driven prevention.
PdM
Architecture
Roadmap
Delivered
MANUFACTURING · RADAWAY
Making email-based order processing reliable with LLMs
Staff were manually reading customer emails, identifying products, and entering orders by hand. We turned a promising AI prototype into a production system that handles it end to end, across languages, formats, and attachments.
-90%
Manual intervention
95%+
Match accuracy
LOGISTICS · FR. MEYER’S SOHN
Eliminating manual data extraction from thousands of daily shipping emails
Operations staff were manually reading German and English logistics emails to pull out routing and scheduling data, every single day. We built an AI pipeline that extracts, structures, and delivers the data automatically.
–80%
Manual effort
2 langs
DE & EN
On-prem
Deployed