Visible Scale

Projects · Medical device distribution case study

An AI system for a medical device distributor.

A case study from a client project: a distributor with SAP Business One, orders by email and field sales in the clinics. We spent two weeks measuring where the time goes and designed a system from it. You can try the approval of an order yourself on this page.

Starting point

Where the time goes.

Four figures from the analysis, from the data in SAP and from the conversations with the team. Everything else builds on them.

Found in the analysis

  • about 17

    orders a day, entered into SAP by hand in the back office.

  • one in four

    orders with a backorder. Finding a substitute takes 10 to 20 minutes.

  • about 30 min

    of typing to record a sales opportunity in SAP after a clinic visit.

  • 3 to 5

    complaints a month, each taking 60 to 120 minutes across eleven steps.

The approach

  1. 1

    Listen and measure

    Workshop with the team, interviews per department, real data from the system.

  2. 2

    Sort the building blocks

    Every recurring task placed by impact and effort.

  3. 3

    Build the foundation once

    Connection to the existing programs, approval interface, audit log.

  4. 4

    Block by block

    Every further building block reuses the foundation and costs less as a result.

The system

A layer between the inbox and SAP.

The system replaces no software. It reads what comes in, fetches the data from SAP Business One, prepares the task and puts it in front of the team for approval.

What comes in

  • Order email

    from clinics and dealers

  • Order as PDF

    attached to the email

  • Field sales dictation

    after the clinic visit

  • Complaint

    with deadlines under the MDR

What the system does

  • Classify

    who is writing, what it is about

  • Fetch data from SAP

    customer, items, stock, read only

  • Write the draft

    in the fields of the target program

  • Check against fixed rules

    required fields, value limits, data protection

Where it goes back

  • SAP Business One

    Customers, items, stock

  • Outlook

    Mailboxes, reply, sending

  • Teams

    Notice to the team

  • Phone app

    Confirmation in field sales

In between, a person approves.Nothing is created without approval.

SAP Business One, Outlook and Teams stay. The system reads from them and writes back, without migration.

One task in detail

An order by email becomes a sales order in SAP.

About 17 orders a day arrive by email, usually with the order attached as a PDF. Today the back office types each of them into SAP by hand. In the design, the system prepares the sales order. This is the interface where your team checks and approves tasks.

Your approval interface

Try it out

Order draft

Hospital, ward 4B

Customer no. K-10482, matched via the sender address
ItemQuantity
Catheter set size MItem no. KS-M-020 · Stock 1414 now, 6 to follow20 pcs
Sterile dressing setItem no. VS-ST-010 · Stock 12010 pcs
Created in SAP after approval

Approval

3 open

Open tasks

Every task is only created in SAP after your approval.

Done

Nothing approved yet today

Sample data: email, clinic and items are made up.

 

Seven steps

From the email to the sales order in SAP. Each step shows above where it happens.

Behind the scenes

The system handles six steps, a person handles one. Only the last one writes to SAP, and only after approval.

The building blocks

Seven building blocks, in this order.

The order is the first. Every further task from the analysis is placed by impact and effort and has its slot in the roadmap. This is the design; what gets built is the client's decision.

Seven building blocks by impact and effort

Start top left. The numbers are the order in the roadmap.

  1. 1Foundation and ordersThe order email becomes a draft sales order in SAP, the back office confirms.
  2. 2Voice in field salesDictate after the visit instead of typing, nothing is created before confirmation.
  3. 3Master data and order confirmationsClean up the data, then order confirmations go out automatically.
  4. 4BackordersSuggest suitable substitutes from the same product family.
  5. 5Sales prioritisationRank open opportunities by activity, value and standstill.
  6. 6Clinic knowledgeEnrich clinic data from public sources, every fact with its source.
  7. 7ComplaintsGuide the eleven steps of the procedure, with translation for the manufacturer.

Roadmap over eight months

The foundation comes with the first building block. A recommendation from the analysis, not a fixed schedule.

The architecture behind it

SAP stays as it is. The foundation is built once. According to the forecast from the analysis, every further building block then costs about 30 percent less than a standalone build.

  1. Existing systemSAP Business One, no migration
  2. ConnectionReads from and writes to SAP
  3. AIAI model in the EU, data-protection filter in front
  4. WorkflowsOn a server in Germany
  5. ChannelsOutlook mailboxes, phone app, Teams
Approval interface

Back office, field sales and quality management confirm every step that creates or sends something.

A similar process in your business?

In the intro call we look at whether a system like this is worth it for you and where we would start. The call takes 30 minutes and is non-binding.

More projects

What else we have built.

A client project that has been in operation since July 2026, and our own tool for the analysis. A click shows the starting point, the flow and the outcome.

Honest

We're transparent. We hold real client logos until explicit clearance. In the intro call we show you relevant examples and, on request, connect you to a reference.