Peter Ratcliffe
Operations, commercial strategy and business development, from scaling established organisations to building new ventures. Has worked with major corporations on new technology. Now finds and implements practical AI.
We build AI that does the work.
Working AI systems, built around real businesses. From first prototype to production.
Commerce
A new collection arrives. The system identifies every piece, assembles outfits and generates them on your model. Your merchandiser approves each one.
Lookbook · FW26
Running
- Collection in6
- Identify0
- Assemble0
- Generate0
- Approve0
Working demonstration on Grand Le Mar FW26 garments and photography. Looks generated by our pipeline from one real model shot. Click any item to look at it.
Five capabilities. One way of working.
These are working demonstrations, not client case studies. Each one runs a whole job from start to finish, and a person signs off the result. Pick one to watch it run.
Built in house by the Blithfield team.
We design, build and run every system ourselves, from the first meeting to production. Nothing is handed off to a third party.
Paddy Archer
Engineering, product development and commercialisation across consumer, industrial and technology markets. Has taken products from concept into the hands of thousands of customers. Now applies AI in real commercial operations.
Our team has worked with
The questions you should ask us.
You are trusting us with your processes and your data. Here is how that works.
- Who builds it?
- Everything is built in house by the Blithfield team.
- How quickly will something work?
- You see a working prototype on your own data within three weeks, running alongside your business rather than inside it.
- What happens after the first build?
- If it earns its place, we take it to production, connect it to your systems and support it. You decide once you have seen it working.
- Who owns the software and the data?
- You do. The code, the outputs and your data stay yours, with a full handover. No lock-in.
- Will it work with our existing systems?
- It works with what you already run: email, spreadsheets, your order, accounts or product systems. Nothing touches a live system until you have approved it.
Three weeks. Three things you can open.
Each week ends with something on your screen, not a slide about it.
Process map
Week 1 · UnderstandWhere the hours go
Order entry34 hrs/wk
Customer queries22 hrs/wk
Supplier paperwork16 hrs/wk
Product content11 hrs/wk
Ranked opportunities
Week 2 · FindScored on value and feasibility
Customer queries92
Order entry81
Supplier paperwork66
Product content49
Working software
Week 3 · ProveRunning on your data
C-901 RachelApproved
C-902 TomApproved
C-903 DanHeld
C-904 AmiraApproved
Illustrative figures.
Measured while it runs.
These numbers come from the demonstration above as it runs. On your data, they are the measures we agree at the start and check at handover.
0Looks processedPipeline count · Commerce demo
—Approved first timeApproval log · Commerce demo
0Held for a personFailed checks · Commerce demo
—Median time to a decisionRun timings · Commerce demo
Bring us the process that eats your week.
Thirty minutes. We will tell you whether AI can take it on, and show you what it would look like running.