Private Limited Company · London EC1V 2NX
Data & AI
Consignment data scattered across telematics, subcontractor emails, and a legacy TMS was consolidated into one live operational picture, with ETA predictions replacing phone calls to drivers.
The situation
Customer service could not answer 'where is my delivery' without ringing a driver or a subcontractor.
Subcontractor updates arrived as spreadsheet attachments, sometimes a day late.
Three telematics providers each had their own portal and no shared identifiers.
Performance reporting to customers was compiled by hand and always contested.
Our approach
The real blocker was that the same consignment had different references in each system. We built a matching layer to establish one identity per consignment first.
Telematics APIs, TMS exports, and subcontractor spreadsheets all land in one event pipeline, with data-quality checks that flag late or malformed feeds rather than silently ignoring them.
Rather than a generic dashboard, screens were designed for the three questions the operation actually asks: what is late, what is at risk, and who needs telling.
A model using historical run times, traffic, and delivery-window data produces ETAs with a confidence band. We deliberately kept it explainable so planners would trust and use it.
The outcome
Customer service resolves the majority of tracking enquiries from the control tower without contacting a driver. Contested performance reports largely disappeared once customers received the same figures from the same source. Planners now intervene on at-risk consignments before the delivery window is missed rather than afterwards.
Related work
A finance team spending eight days a month matching settlement files by hand moved to an automated reconciliation engine with a clear exception queue — closing the month in under a day.
Month-end close time
Transactions auto-matched
Adjustments with audit trail
Read the case study
Three ageing storefronts sharing one overloaded template were rebuilt as a headless front end on a shared component library, cutting page load times and ending the annual Black Friday panic.
Mobile LCP, product pages
Mobile conversion rate
Annual hosting cost
Read the case study
Referral, intake, and consent handled on paper and by phone became a patient portal with role-based clinical access, cutting administrative handling time per patient by more than half.
Admin time per new patient
Appointment no-shows
Intake completed before arrival
Read the case study
Next step
Send a short description of the problem and we will reply within one business day. The first call is a 30–45 minute conversation with someone senior — free, and with no obligation.
sales@interwebservices.co.uk128 City Road, London, EC1V 2NX, United Kingdom