Sixteen problems we keep solving on repeat — each one a sample of what we can do. If one sounds like your Tuesday, that’s the conversation.
Sixteen examples of what we can do
Workflow & Process Automation
Orders arrive by email and someone reads, checks, and retypes every one before it reaches the ERP.
See the capabilityAI Extraction & Validation
Invoices, certificates and orders arrive as PDFs. Someone reads each one — and errors surface weeks later.
See the capabilityReporting & Power BI
The weekly report is rebuilt by hand, and finance and sales still bring different numbers to the meeting.
See the capabilityForecasting & Decision Support
The forecast lives in one spreadsheet and one person’s head. Ask twice, get two answers.
See the capabilitySoftware & Platform Engineering
The business runs on a spreadsheet nobody dares touch — or a legacy system nobody dares change.
See the capabilitySystem Integration & ERP
The systems exist, but the data still moves by export, reformat, and manual import.
See the capabilityComputer Vision
A trained eye reviews drawings, photos and samples all day — but the answer never becomes structured data.
See the capabilityVision in Production
The demo worked. The line is harder — lighting shifts, hardware drifts, edge cases arrive daily.
See the capabilityProduct Data AI Assistant
The answer to a product question exists across ERP, PDFs and old emails. Nobody knows where.
See the capabilityQuote Automation for Job-Shops
Every RFQ is a pile of drawings, files and notes — and quoting starts from a blank screen.
See the capabilityReal-Time Production Views
The ERP holds the data, but the floor still runs on whiteboards, spreadsheets and side conversations.
See the capabilityAI Quantity Takeoff
Drawings are measured and counted by hand — and the number is separated from the visual source.
See the capabilitySensitive Registers Out of Excel
A sensitive register grew up in Excel — more people, more fields, more risk, no controls.
See the capabilityOne Connected Operational Flow
Work happens through calls, paper and whiteboards. The system only sees the end result.
See the capabilityContract Register with Index Flags
The contract exists, but the date that matters is buried inside it — and passes quietly.
See the capabilityRequirements to Ranked Shortlist
Three lines of requirements arrive, and the shortlist still gets built by hand without visible reasoning.
See the capabilityEight anonymised engagements
Built, scoped or proven in real dialogues — written up without names. Each one is a pattern we can apply to your operation.
Make your product data AI-readable — answers with a source, never a guess
An organisation with ~200,000 products behind clean APIs found that AI tools hallucinated, missed fields or built shadow copies of the database. We built an orchestration layer where AI never talks to the data directly — natural language becomes a schema-validated query, and every answer is traceable to a source page.
Industry product databases & large cataloguesRead the capabilityQuote automation for job-shop manufacturing — your CAD files already hold the answer
A request lands as an email: a PDF drawing, sometimes a STEP file, quantities in free text. Someone reads it, calculates in Excel and re-keys it into the ERP. We iterated a concept through three versions — and learned that deterministic CAD data beats AI interpretation, cutting the estimate from 240–270 hours to roughly 40.
Laser cutting & contract manufacturingRead the capabilityReal-time production views on top of your ERP — replace paper orders without replacing the system
Orders printed on A4 and sorted into trays. Assembly times nobody measures. Multi-level product structures forcing operators to report in exact sequence. We design visualisation layers on top of the existing ERP — four connected views that replace paper, capture real times as a by-product, and never touch the ERP itself.
Make-to-order manufacturing & assemblyRead the capabilityAI quantity takeoff from construction drawings — measured by the machine, approved by your sales team
Eight salespeople scale-measuring facade drawings in PDF tools — or on printed A3 with pen and paper — roughly 1,400 quotes a year, where only ~20% become orders. We apply a vision pipeline that reads the full drawing package and produces a quantity list with confidence scores and traceability, for the salesperson to review instead of measure.
Building products & construction supplyRead the capabilitySensitive registers out of Excel — GDPR as a property of the system, not something to remember
A 180-person consultancy ran its landowner process in Excel: personal IDs, addresses and bank details in the same file, GDPR masking done by hand "when it happens", and final deliveries glued together into PDFs one full working day at a time. We design platforms that own the data lifecycle — register, status, delivery log, and automatic retention.
Infrastructure & field-operations consultanciesRead the capabilityFrom whiteboard, paper and phone calls to one connected flow — order to invoice
A recycling operator ran ~600 transport orders and ~350 invoices a month on a whiteboard, paper order slips and phone-reported weights — with two days a month spent re-keying everything into the accounting system. We build operational platforms where the order is the system: customer → order → driver → weight → invoice, with the article register as the pricing engine.
Recycling, logistics & field-service operationsRead the capabilityA contract register that flags in time — stop losing the index increases you are entitled to
Consultancies billing by the hour against public framework agreements are entitled to annual index adjustments — but the process runs on PDFs in folders, metadata in Excel and someone remembering to start. At one 40-consultant firm a missed adjustment meant a 17.6% uplift was never claimed. We design contract registers with AI parsing, deadline watch and an index calculation engine.
Professional-services firms on framework agreementsRead the capabilityFrom three lines of requirements to a ranked shortlist in minutes — with evidence, not guesses
A working proof of concept: free-text role requirements become an editable ideal profile, hard requirements are enforced by code (never by AI), candidates are ranked 0–100 by semantic match, and every screening question is answered with evidence from the CV — or an explicit "no evidence". Built end-to-end in about a month.
Recruitment, staffing & consultancy salesRead the capabilityEvery example is anonymised from a real engagement — and the systems are still in use.
How do we know which of these applies to us?
Look for work that is repeated, manual, and already written down somewhere — a spreadsheet kept by hand, a report rebuilt every month, a check done by eye, a quote priced from drawings. Every capability on this page started as one of those. If a task happens often, takes roughly the same shape each time, and still needs a person to keep it moving, it is worth a conversation. If it happens rarely, or looks different every time, it usually is not.