Two parallel AI programmes and continuous ERP development, running against live production environments.
Client identities and commercial terms are withheld on this page. Scope, workstreams, architecture and milestone structure are published with the customers' permission, so that prospective clients can see the shape of the work we take on.
A six-workstream AI programme for a leading manufacturer of engineered pipe supports and cryogenic products, delivered as one common AI platform rather than six separate tools.
| No. | AI workstream | Primary objective |
|---|---|---|
| 1 | AI for Spring Support — Pre-Order Engineering | Read customer PDF / Excel requirements; compare with the manufacturer's catalogue, global standards and historical data; identify deviations and generate standard engineering outputs. |
| 2 | AI for Spring Support — Post-Order Engineering | Validate post-order engineering information against approved standards and reference data; identify deviations and generate the required engineering outputs. |
| 3 | AI for Cryogenic — Pre-Order Engineering | Read customer PDF / Excel requirements; compare against approved references and generate specification, material, deviation and drawing data outputs. |
| 4 | AI for Cryogenic — Post-Order Engineering | Validate post-order cryogenic engineering data against the applicable standards and reference information. |
| 5 | AI for Engineering QC | Validate engineering requirements and outputs against approved standards after post-order engineering, and identify deviations and mismatches. |
| 6 | AI for QA Documentation | Generate and support QA documentation using applicable historical reference documents and approved documentation structures. |
Workstreams 1 to 4 share a single processing pipeline. Only the reference sources and the output templates differ.
| Standard output | Purpose | Format |
|---|---|---|
| Material Data Sheet | Structured material information extracted or derived from the customer requirement and reference data | Standard Excel |
| Deviation Data Sheet | Customer requirement mismatches and deviations against the applicable standard or reference | Standard Excel |
| Drawing Data Sheet | Drawing and specification information required by the agreed engineering output | Standard Excel |
The Engineering QC workstream validates applicable engineering requirements and post-order engineering information against approved standards and reference information, identifies mismatches and deviations, and presents the findings for engineering review.
The QA Documentation workstream uses approved historical reference documents and agreed document structures to assist in preparing the required QA documentation, with the final document list based on the examples supplied by the customer.
Because all six workstreams are developed simultaneously, milestones are not assigned to individual projects. Each milestone represents a measurable phase of the overall programme and covers progress across all applicable workstreams. Acceptance is based on demonstrable software capability and agreed deliverables — not merely the passage of a calendar month.
Eight AI initiatives layered into a live multi-service logistics platform covering Container Freight Station, customs clearance, freight forwarding and transportation.
The objective is not to create a separate AI application. Each capability is delivered as a functional layer inside the existing web and mobile applications, so users reach AI from the workflow they are already in. The initiatives are implemented progressively, prioritising those that provide the highest immediate business value.
Management and users interact with the platform using natural-language prompts instead of navigating multiple screens and report filters. The AI understands the request, retrieves the data, processes it and returns a summary, table, chart, graph, PDF or Excel file.
Document extraction across the different customer and vendor invoice types used in import and export customs clearance, converting unstructured document data into structured records for the platform.
Extraction and processing of ICEGATE-related documents and data for CFS import and export operations, combining API-based retrieval, document processing and AI interpretation where applicable.
A user supplies a rough description and the AI returns a properly structured business email in the required communication format — standardising customer communication and cutting the time spent on routine correspondence.
A natural-language support interface using authorised platform data and configured business processes, so customers and internal users can self-serve routine queries about container location, shipment status and required documents.
Daily status reports generated automatically in the customer's own reporting format from the latest operational data, and distributed through the configured communication channels.
A document assistant that reads SOPs, process flows and manuals — including collections running to hundreds of pages — and answers process questions with indexing and retrieval instead of manual searching.
A quotation assistant that analyses internal business data, historical information, customer requirements and external information to help prepare competitive quotations from an emailed requirement.
All eight capabilities are integrated into the existing web and mobile applications rather than delivered as standalone tools. Users access AI directly from their existing workflows, with the platform's own authorisation rules applied to every request.
Alongside the two AI programmes, both ERP platforms are under active development in live production environments.
Gate in / gate out pass generation, approval dashboards, container entry and print flows across the import and export lines.
Multi-level leave approval routing, attendance management, employee evaluation, satisfaction survey and role-based dashboards.
Public data entry, security admin approval, QR-based check-in, signed one-click email approval and visitor photo prefill.
If your operation has documents being re-typed, reports being assembled by hand or standards being checked manually, there is usually a workstream in it.