--- url: https://l-be.com/ai-case-study canonical: https://l-be.com/case-study site_tree: https://l-be.com/ai#tree whole_site: https://l-be.com/ai-all language: he updated: 2026-10-08 ---
Case study
Ida's Home
The whole business in one process, from sale to payment collection. And above it, an AI layer connected to every department.
[Image: Illustration: a central cube connected by dashed lines to six modules around it]
About the company
Ida's Home designs and sells home furniture: made-to-order sofas and armchairs, lighting, rugs and textiles. Alongside the products it provides planning, installation and design consulting services. The company has 30 employees. It sells on its website, in the factory store and in the showroom, and works with several warehouses, a carpentry workshop and many suppliers.
The challenge
Every order at Ida's Home is a small project:
- A product made to measure and in the fabric the customer chose.
- A supplier that has to meet deadlines.
- A warehouse that receives and picks.
- A delivery that has to be coordinated.
- A customer who is waiting, and a payment that has to be collected.
When the information is scattered across people, files and systems, the team is busy chasing. Where is the order? When does the supplier deliver? What's in stock? Who has paid?
A sale takes time, because it's hard to find the right product. A new salesperson needs months to learn the job. And management doesn't see one picture it can rely on.
Management set a simple test:
If the team is still chasing customers, statuses, deliveries, price lists and documents — the project has failed.
The approach
How we approached it
Specification before building
We sat with management and went through every process, until it was clear how it works in practice.
One source for every piece of data
Every system is updated from the others automatically. No double entry and no "versions".
One process, all the way through
First it works in the field, and only then do we expand to the next process.
Automation with human approval
Automation where the rule is clear. A person approves where there is money or a customer involved.
What was built
The core processes of the business, connected end to end
Sales in every channel
- Website, factory store, showroom and points of sale.
- A dedicated point of sale (POS) with credit card processing and payment by link.
- Every order, from every channel, enters the same track.
One customer, one order and one picture, no matter where they bought.
Purchasing and suppliers
- Purchase orders go out to suppliers automatically, according to what the customer ordered.
- Every supplier regularly receives a picture of what is open with them.
No order is forgotten on the way to the supplier.
SLA with suppliers and with customers
- Every product has a calculated delivery date, and everything that goes past it is tracked.
At any moment you can see what is delayed and with whom, before the customer calls.
Inventory and warehouses
- Inventory management across several warehouses.
- A product that is already in stock is allocated to an order automatically.
You sell what you have, and don't reorder what is already on the shelf.
Smart scanners
- Scanning on a phone at every stage: receiving from a supplier, picking, dispatch for delivery, returns and repairs.
- Every scan moves the order forward automatically.
The warehouse updates the whole company as it works, without typing.
Shipping and delivery coordination
- The customer sets a time themselves, through a link.
- The couriers are assigned to a route.
- The courier app documents every delivery with a photo and a signature.
- An interface for managing the couriers and paying them.
Every delivery is coordinated and documented from the start to the handover.
Marketing and customer service
- Automatic messages to the customer at every stage of the order.
- A proactive update to the customer, and a request for feedback at the end.
The customer knows what is happening without asking, and the service team is free to handle what really needs handling.
Complementary products
- An offer to the customer to complete their order, with a personal link.
The basket grows without manual work.
Payment collection and accounting
- Sync with accounting.
- Credit card processing.
- Collection tracking, and detection of gaps between the systems.
A true picture of the debt, not a number where you have to guess what is behind it.
BI and management
- Dashboards, cash flow and insights that update themselves.
Management sees the business in real time and decides based on data.
The AI layer
Above all these processes stands an AI system connected to the data of every department.
- An employee asks in plain Hebrew, for example "Which orders are delayed at a supplier?", and gets an answer based on the company's data.
- A useful report is saved, and can be run again with one click.
- A manager asks for a dashboard in words, and the system builds it.
- Behind every number stands one agreed business definition, so everyone is talking about the same "sold" and the same "late".
Responsible AI. An AI system connected to a company's data has to work within clear boundaries. The control is built in:
- The AI suggests, and a person approves.
- Every employee sees only what their role allows.
- Sign-in uses two-step verification.
- Customers' personal details are encrypted, so that they do not pass to the AI or leave the company.
- Every action is logged.
The result
What changed
- One continuous process From sale to payment collection, instead of information scattered across people and files.
- Updates from the field The warehouse, the couriers and sales update the system as they work, from their phones.
- One picture for the whole company Every department sees the same order, in the same status, at the same time.
- A reliable picture for management Of orders, inventory and debts. The system exposed gaps that had not been visible before, and they were dealt with.
My role
- Specifying the processes with management.
- Setting up the CRM from scratch.
- Connecting all the systems to each other.
- Building the AI layer and the apps.
- Implementation on the ground.
- Ongoing support.