Skip to content

For care providers, hospitals and insurers

Twenty years of system landscape.One data foundation.

Your infrastructure grew faster than your documentation. We build the layer underneath — a data crawler that runs every night through everything that has a database. What you build on top is up to you.

The actual gap

Nobody designed it this way. It simply happened.

No hospital group with two thousand people drew its system landscape on a whiteboard. It grew — through acquisitions, through projects, through people who left long ago. Every single system was the right decision at the time. Together they do not add up to a data foundation.

Legacy

Systems nobody touches anymore

They run, they are critical, and the colleague who set them up left years ago. Nobody wants to go near them.

Acquisitions

Landscapes that never merged

Two ERPs, three ticket systems, four file shares. Integration was on the list every year and postponed every year.

Evidence

Audits as archaeology

Six weeks before the date, the gathering starts. From mailboxes, spreadsheets, repositories. By hand.

What we do

The data crawler.

A process that runs every night through your entire estate and makes it queryable — line-of-business systems, billing, accounting, CRM, personnel data, source code. Read-only, fully logged, without touching a single running system.

Every night

Current, not once a year

The estate is captured daily while your systems are idle. By morning it is up to date.

Complete

Including where nobody usually looks

Discrepancies between documented need and billed service, the link between a commit and a ticket, histories no report shows.

Untouched

No system is modified

Read-only throughout. Switch the crawler off and everything runs exactly as before.

Protected

Encrypted, separated, attested

At rest with AES-256-GCM, in transit through a WireGuard tunnel. Operated in an own data centre in Germany, attested under BSI C5 Type 2.

Connected

What we already read.

Not a list of what would be possible. This is running.

Line of businessMedifox · CuraSoft
Billing and accountingNoventi · DATEV
CRM and personnelHubSpot · Personio
Source codeGitLab · GitHub

In trial, not yet in regular operation: Atacama · CGM KIS.

The names above are examples. Anything with a database can be connected — read in full, kept over time, stored so it can be queried. Whether we know a system is beside the point. Whether it has a database is not.

Source code is indexed via SCIP. Providers between 25 and 5,000 employees, one to 41 locations. First source connected: two to three days.

What you build on it

From then on, you ask the questions.

Your own agents, your own analyses, your own criteria — with Claude or any other model. The questions below are not hypothetical. They were asked and answered.

From line-of-business, billing and personnel

“Which patients have a care level that no longer matches their documented need?”
Medifox · Noventi
“Where does documentation quality deviate systematically — and on which topics?”
CuraSoft · Personio
“What are delayed payments waiting on, and with which payer?”
Noventi · DATEV · HubSpot

From source code and repository

“Which changes in the last 90 days touch regulated functions?”
GitLab · SCIP
“What depends on this function — and what breaks if we change it?”
GitHub · SCIP

We start with a single system.

Medifox and DATEV were connected and running after three days. In the first conversation we walk through your landscape and show concretely which sources can be connected and what becomes answerable afterwards. No preparation needed on your side.