It’s been great to work with the Unloq team. Their knowledge of AI is vast and they’ve helped us deliver AI solutions that will improve customer engagement and give better insight.
The results
5,000+
custom data fields audited
8 weeks
from idea to live delivery for XPENG’s UK launch
599
live support tickets analysed to pinpoint quality gaps
IM Group had valuable data across multiple disconnected systems but no unified platform to bring it together, no structured pipeline architecture, and no way to surface intelligence at speed.
Customer, dealer, and vehicle data sat in separate platforms with no integrated layer connecting them. Reporting was manual and siloed. The XPENG brand launch needed a production-grade AI customer service capability, but the underlying data infrastructure did not exist.
We worked across four workstreams: data architecture, pipeline engineering, AI delivery, and platform build. Each one built on the last.
The XPENG UK build went from idea to live delivery in eight weeks, and it left IM Group with a reusable data and AI foundation for future reporting, commercial intelligence and operational use.
How we did it
- 01
Audit the data
We went through IM Group’s data across CDP, HubSpot, MVRIS and Azure, and reviewed more than 5,000 custom fields to find out what was there and what could be trusted.
- 02
Design one layer
We designed the target architecture on Microsoft Fabric with a medallion structure, so customer, dealer and vehicle data can sit in one integrated layer.
- 03
Build for launch day
We built the pipelines and data infrastructure under an AI customer service bot for XPENG’s UK launch, using 599 live support tickets to pinpoint quality gaps and testing it against safety-critical checks before it went live.