The work, labelled for what it is.

The problem

Nobody buys a sofa from a web page. They shortlist on one, and a showroom that cannot be shortlisted online loses the visit before it happens.

The customer who matters most is often a designer or architect specifying a whole job. That person needs to gather pieces from many design houses and send one considered request, not nine separate enquiries.

What was built

• A catalogue of thousands of pieces from more than sixty design houses, mainly European.
• A real hierarchy: by room, by type of piece, by design house and by family, so one piece is never listed five times.
• Filtering by room, brand and availability.
• A quote list, so a specifier sends one request for a whole job.
• A separate contract library for restaurants, hotels and offices.
• Online purchasing.

How it was made

AI tooling did the drafting. Every judgement that mattered stayed human, and the owner catching a chair filed as a sofa, or a photograph that was not the piece, is part of the process rather than a failure of it.

Full disclosure: Ivan, who runs MM8, co-owns Le Belle Arti. It is not an arm's length client, and the studio lives with every decision it made there, every day.

No sales figures are claimed here. What can be checked is that it is built, running, and held in the business's own accounts.

The problem

A sale in a furniture showroom passes through many hands: a quote, a deposit, an order to the supplier, a confirmation, a delivery. It lived in a paper binder and a spreadsheet, and the certainty of each was worth keeping.

What it does

• A pipeline from lead to delivery.
• Quoting, sales orders and purchasing.
• Invoice tracking and inventory.
• Roles, so each person sees their part of the job.
• Discount approvals.
• An audit trail of every change.

Where it stands

In development, and not yet in daily use. The picture above is a screenshot of a practice copy of Orli, filled with invented clients.