India / USA · Industrial IoT / Facilities Management
Helix Sense
“The sensors were everywhere. The insights were nowhere.”
Sensors deployed but zero data processed - turned into real-time IoT intelligence across 4 operational domains.
Key Results
4 domains
Live operational data
Was zero
Real-time
Sensor processing
Databricks streaming pipeline
NLQ live
Copilot querying
Plain English across all 4 domains
Full ROI
Hardware investment
Was zero operational return
Tech Stack
The Situation
Helix Sense had done the hard part - at least, the part that most people think is the hard part. They had deployed sensor hardware across multiple facilities: vibration sensors on critical equipment, energy meters on every major circuit, environmental monitors across buildings, and asset tracking across their estate. Significant capital expenditure. Physical installation complete. But when the operations team asked "what is the energy consumption trend on Building C?" - there was no answer. The sensors were transmitting data. That data was going somewhere. But there was no pipeline to catch it, no processing layer to clean and structure it, and no visualisation layer to surface it. The hardware investment had delivered zero operational return. The team was flying blind with expensive sensing equipment attached to everything.
For industrial IoT, facilities management, and manufacturing companies, this is a surprisingly common situation:
- ✓
You've deployed sensors or IoT hardware but the data isn't connected to anything useful
- ✓
Your engineers can see raw telemetry in a vendor portal but operations managers can't act on it
- ✓
You have data from equipment, energy, and building systems - but it's in three different vendor tools that don't talk to each other
- ✓
Leadership approved the IoT investment on the promise of insights, but those insights haven't materialised
- ✓
Your data scientists or analytics team says "we could do a lot with that sensor data - if we could access it"
- ✓
"We have all this data but we don't know what to do with it" - the most common sentence in industrial IoT projects
If three or more of these describe your operation, you're looking at the right case study.
The Root Problem
- 1
Sensors were deployed and transmitting across multiple facilities - but no data processing pipeline existed to handle the output
- 2
Zero visibility for operational teams: Engineering, Energy, Asset Management, and Facility Management were all flying blind
- 3
Sensor data from different vendors and domains was in incompatible formats with no common data model
- 4
Significant hardware capex had been approved on the promise of operational intelligence - with no ROI materialised
- 5
Business users had no way to query or explore sensor data - everything required direct engineering involvement
How We Fixed It
Ingest all sensor streams into a single processing layer
The first task was standing up a Databricks ingestion layer that could accept data from all sensor sources - regardless of protocol, format, or vendor. Azure Event Hubs was used as the streaming buffer, with Databricks Structured Streaming consuming and processing the data in near real-time. All raw sensor data now lands in Delta Lake with full lineage and schema enforcement.
Design four operational domain models
Rather than building one monolithic data model, we designed four separate domain models - one per operational domain: Engineering (equipment telemetry, vibration, temperature), Energy (consumption by circuit, efficiency ratios, anomaly detection), Asset Management (condition scoring, maintenance signals, age-based risk), and Facility Management (environmental conditions, occupancy patterns, space utilisation). Each domain has its own semantic layer and serves different stakeholders.
Build Power BI dashboards per operational domain
Each of the four domains got its own Power BI dashboard designed specifically for the operational audience that uses it. Engineering dashboards show equipment health scores and anomaly alerts. Energy dashboards show consumption trends, efficiency benchmarks, and carbon signals. Asset Management shows condition scoring and maintenance priority queues. Facility Management shows real-time environmental conditions and occupancy trends.
Deploy Copilot for natural language querying
Microsoft Copilot was integrated across all four domains, allowing facility managers, engineers, and operations leads to ask questions in plain English - "Which HVAC units have shown vibration anomalies this week?" or "What was our energy intensity per square metre last month vs. the same period last year?" - and get answers drawn directly from live sensor data. No query language. No analyst needed.
Measured Outcomes
Operational data domains live
0
4 (Engineering, Energy, Assets, Facility)
↑ Key win
Processing pipeline
None - data going nowhere
Databricks real-time streaming
↑ Key win
Team visibility
Zero across all domains
Real-time dashboards per domain
NLQ data access
None
Copilot across all 4 domains
IoT hardware ROI
Zero operational value
Full operational intelligence
What This Means For You
What this means for IoT and facilities management companies
The most common reason IoT investments fail to deliver value is not the hardware - it's the missing processing and analytics layer between the sensors and the people who need to act on the data. The sensors work. The data is there. The gap is the pipeline that transforms raw telemetry into operational insight. This gap is fixable in a matter of weeks, not months. If your sensors are transmitting and your team still can't answer basic operational questions from that data, the problem is solvable - and the answer almost certainly involves Databricks or Microsoft Fabric as the processing layer, with Power BI on top for the visualisation your operational teams will actually use.
Next Step
Is this your situation?
Book a 30-minute call. No slides, no pitch. We'll look at your specific setup, tell you what's causing the problem, and what a realistic fix looks like - including timeline and cost range.