The data foundation
What the Newtack data foundation is built from.
Your EHR, your billing system, and your payroll system each do their own job well. None of them was built to answer a question that spans more than one, such as whether a program that is short-staffed this month is about to miss a billing target next month. That kind of question falls into the gap between your systems, and today it gets answered by hand, late, or on instinct.
The Newtack data foundation closes that gap. It reads every system your organization runs, the EHR among them, and the business functions those systems support, such as billing, payroll, and care planning. What comes out is understanding and decision support, held in one place your team can act from. The action itself still happens inside each system, the same way it does today. Automating that work is something to take up later.
Leadership gets a place to look before a rate negotiation, a board meeting, or a staffing decision, with one agreed number behind every measure.
In use
What leadership uses it for.
For example, planned care hours come from your care-planning system, and delivered care hours come from your EHR. On their own, each system shows only its half. In the data foundation, both numbers sit in the same place, broken out by location and clinician, so a staffing gap shows up before visits are missed. A finance leader can ask it for revenue by program ahead of a payer meeting, and a program director can ask it which locations are behind on signing notes. All three get an answer built from the same agreed numbers.
The parts
What it is built from.
The data foundation is built from several layers, stacked in order between your systems and the reports your team reads. Each layer works the same way across the EHR, billing, payroll, care planning, and every other business system you run, and each layer has one job. Most of that software is released under what is called a permissive license: if the company behind it ever stops supporting it, anyone, including another company, can keep developing the code. Your tables are also kept as files in an open format that any major database can read. That is why replacing one layer's software later does not mean rebuilding the layers around it, or rebuilding the reports and the agreed definitions your team already has.
Newtack runs and watches every layer, so your team does not have to learn how any of this works underneath. Here is what each layer does.
Scheduling and run tracking
Runs every update to your data in the right order, on schedule, and keeps a record of each run, across every system it reads from.Runs on Dagster, open source.
Storing your system logins
Holds the login for each of your systems and uses one only at the moment a job needs it. You can revoke access in that system at any time.Runs on secrets vault software.
Copying your records
Copies new records out of your systems on a schedule, without creating duplicates or leaving anything out.Runs on dlt, open source.
Building your reporting tables
Turns the copied records into the tables your dashboards and reports read from, and checks the numbers before anyone sees them.Runs on dbt, open source.
Storing the tables
Keeps your tables as files in an open format that any major database can read, and answers each query from a database engine sized to your data.Open file formats and a database.
Documents, calls, and recordings
Turns contracts, recordings, and video into text your team can search, limited to what each person is already allowed to open.Newtack's own software.
Agreed definitions, and who sees what
Holds one agreed definition for each measure, such as census or revenue, drawn from whichever systems feed it, and the rules on who may see which records, so every view returns the same number.Runs on Cube, open source.
Dashboards, questions, and other tools
Shows the data to your team, and to any other software that needs it: reports, dashboards, plain-language questions, and anything you add later.SQL or REST interfaces.
Durability
Why a vendor change does not strand you.
Two of the companies behind software in these layers changed hands within two months of each other in 2026, and both changes were public. Prefect announced in July 2026 that it was acquiring Dagster Labs, the company behind the layer that schedules and tracks your data updates, and said in its own announcement that the product keeps its name, its open license, and that no deployment needed to be rebuilt. In a separate deal, Fivetran and dbt Labs, the company behind the layer that builds your reporting tables, completed a merger, and dbt Labs published that the tool keeps the same open license it has always had. Neither change required anything in your data foundation to be rebuilt.
Next step
Talk with us and see the demo.
If your technical team wants the detail behind any of this, the next page goes through the data foundation one layer at a time, with the license and vendor history for each layer. Write to [email protected] to set up a conversation. We will share a screen and walk you through the demo. The demo runs on sample data, the same way the product works.