Client names are withheld where confidentiality requires it, and we describe what the work actually involved rather than what it would be flattering to claim. Where a project found errors in the numbers an organisation had been relying on, we say so.
A warehouse of more than four billion rows was moved from SQL Server to an Azure Synapse dedicated SQL pool. Distribution and partitioning were redesigned table by table for the MPP engine rather than copied across, and every table was reconciled on counts, checksums and samples before it was accepted. Power BI was then served through an API layer so that reporting no longer competed with the operational system.
More than 1,200 QlikView objects were catalogued against real usage, rationalised, and rebuilt in Power BI on semantic models over an Azure Synapse warehouse. Calculation logic that had accumulated inside the BI tool was lifted into a governed data layer. Every report was validated number by number against QlikView before its wave went live, and the legacy platform stayed available in parallel until each wave was signed off.
Commercial, operational and finance functions ran on separate systems, so what the CRM said had been sold did not agree with what was invoiced. One platform was built covering pipeline, orders, invoicing and finance on a single data model, with the executive dashboard reading that same record rather than a monthly consolidation assembled by hand.
The group ran separate systems per department, which fragmented the patient history exactly where continuity of care depends on it. One electronic patient record was built covering scheduling, bed management and pharmacy, exchanging data over HL7 FHIR with laboratories and existing systems, with role-based access, encryption and audit trails over sensitive health data.
The operator ran on spreadsheets and disconnected third-party trackers, with no single live view of the fleet and no dependable cost figure. One platform was built ingesting embedded IoT telemetry at event grain, joined to fuel and maintenance records, so that consumption, driving behaviour and cost per kilometre became measured rather than estimated.
Sensor data kept at source grain so a specific journey remains investigable.
Cost per kilometre
Measured
Calculated from consumption, fuel and maintenance rather than apportioned.
Maintenance model
Preventive
Scheduled on odometer or hours of use instead of triggered by failure in service.
End to end
7 months
Discovery, ingestion, portal, driver app, pilot acceptance and rollout.
Next step
Arrange a technical review
A 45-minute working session with an engineer rather than a sales call. You describe the problem; we set out what we would build, in what order, and what it is likely to cost.