The problem
Every department behind its own screen
Our solutionOne live view across the business
Seven dashboards covering the overview, work orders, stock, depots, suppliers, sales orders and planning, refreshed from the ERP through the day.
Case studies
AI and automation
One view of the whole business, built on the ERP they already had
A UK manufacturer with an office team and a warehouse ran everything through a mature ERP system. The data was good. Nobody could see across it, chasing suppliers was a daily manual job, and shipments lived in a spreadsheet. In two months we had the first version of an operational intelligence and workflow platform live: dashboards, automated purchase-order chasing and an AI assistant that answers from real numbers. It has kept growing since.
- Client
- UK manufacturer, 20 to 30 people
- Scope
- Dashboards, supplier chasing, shipments, AI assistant
- Timeline
- Started April 2026, still expanding
- Status
- Live, 10+ people across 5 roles
The data was all there. Nobody could see across it.
Manufacturing, stock, sales, purchasing and invoicing each lived behind their own screen. Nobody could tell which customer orders a late supplier delivery would hit, or which components would run short. Routine work, like chasing suppliers to confirm orders or tracking what had shipped, was done by hand.
The problem
No line of sight from a late delivery to a customer order
Our solutionEvery customer order traced back to its supply
Each order is checked line by line against stock, works orders and open purchase orders, down to the components inside each kit. When a supplier delivery slips, the orders it puts at risk show up, and planning lists the shortfalls to buy.
The problem
Chasing suppliers to confirm every order
Our solutionPurchase-order tracking that reads the inbox
Supplier replies are matched to their orders automatically. Chase emails go out each weekday, within limits the team sets.
The problem
Shipments tracked in a spreadsheet
Our solutionA picklist tracker for the warehouse
Every pick list from the ERP, the shipment it is on and when it was collected, with labels and photos.
The problem
Questions that needed a report to answer
Our solutionAn AI assistant that works from live data
Ask which orders are overdue or what is short for next week. It answers from the ERP using the same logic as the dashboards.
Seven dashboards and five tools, in one place
Behind one sign-in, with the AI assistant available on every page.
7 dashboards
- Overview
- The headline numbers across the business, on one page.
- Work orders
- Everything being built, and what is running late.
- Inventory
- Stock levels, usage and stock that is tied up.
- Depot performance
- How each depot is doing.
- Supplier performance
- How reliably each supplier delivers.
- Sales orders
- Every customer order, and whether each line can ship as promised.
- MRP / planning
- Works orders to raise and shortages to buy, down to the parts in each kit.
5 tools
- Purchase order tracker
- Reads supplier replies and chases late confirmations automatically.
- Picklist tracker
- Every pick list, the shipment it is on and when it was collected, with labels and photos.
- Key customer reconciliation
- Checks a key customer's orders against the ERP and flags any differences.
- Work instruction checks
- Uses AI to check that work orders match their standard operating procedures.
- Stock finder
- Finds where any part is held, in seconds.
Inside the platform
Illustrative only, all the data in these illustrations is fake/anonymised.
Sales ordersEvery open order gets a verdict: can it ship as promised? Purchase-order trackerSupplier emails matched to orders. Chases go out automatically. The AI assistantAnswers from the ERP, with links to the records behind it. PlanningWhat to build and what to buy, through every level of each kit. Picklist trackerEvery pick list, its shipment, and when it was collected.
Select a screen to enlarge it
Four calls that shaped the build
- Decision 1
Build it to be added to
The client asked for a platform they could keep building on. Each tool is its own module on a shared base, so new features slot in without reworking what is already live. The picklist tracker and stock finder were both added that way after go-live.
- Decision 2
Keep the ERP, build on top of it
The business already ran on a mature ERP, and there was no reason to replace it. The platform sits on top of its data, reading from a read-only copy of the database, so staff keep working in the system they know and the live system is never touched.
- Decision 3
Keep the maths out of the AI
A deterministic framework. Counts, late-order verdicts and shortages are worked out in code. The AI assistant only explains the result. This makes it give the same answer as the dashboards instead of hallucinating one.
- Decision 4
Test risky changes in the background
When we changed how orders are judged, the new logic ran silently alongside the old and logged what it would have done. The team checked the results before anything went live.
What changed for the business
Before and after each tool went live. Numbers from the client's own reporting.
Supplier response rate
Purchase order tracker with automated email chasing; replies are read by AI to work out each order's status, so staff know what to focus on.
- 60%
- 88%
On-time deliveries
In three months. The same tracker and chasing, with every order's status kept current from supplier replies.
- 67%
- 79%
Rating from their key customer
The effectiveness score their key customer gives them as a supplier. Still climbing each week.
- 78%
- 93%
Queries to senior staff
Sales order tracker and dashboard. The platform pulls the data together and applies the calculations, so staff decide without escalating.
- Baseline
- 60% fewer
£70,000
of ‘lost’ stock found on day one
Stock visible in one place for the first time, measured against what the records said was there.
What it runs at today
- purchase orders in the tracker
- 2,500+
- suppliers across the tracked orders
- 500+
- people across 5 roles, office and warehouse
- 10+
- tools the AI assistant can use to look things up
- 30+
- dashboards and tools in one view
- 7 + 5
Safe to put in front of the whole team
- 01
Inside their own Microsoft cloud
Hosted in the client's existing Microsoft Azure environment, under their control.
- 02
Data handled in line with GDPR
A data strategy that follows GDPR, covering what is held, where it is kept and who can see it.
- 03
Single sign-on
Staff sign in with the work accounts they already have. Easy for them, and access is controlled in one place.
- 04
Least privilege, by role
Role-based access and row-level security in the database mean each person reaches only the parts of the platform their job needs.
- 05
Defence in depth
Layered protection, so a single compromised account can't open up the rest of the platform.
Got data you can't see across?
We say what we would do. Then we do it.