Pragmatic AI

Your clients' data. Handled like it matters.

Your cases hold health, disability, legal and claims information about real people. Here's exactly what we ask for, what we do with it and when it's gone, at every stage of working together.

What happens at each stage.

  1. The free session

    No data changes hands. We talk about how your cases arrive and where the time goes, using examples we make up, not yours.

  2. Looking at a sample

    We ask for a small sample of past cases with names and identifying numbers removed. If that isn't practical, we work inside a workspace you control, or you run the review yourself and share only the results.

  3. Mapping how your cases move

    We sit with your team and watch the work. Notes and maps describe your process, not your clients. Anything that identifies a person stays out of them.

  4. A pilot on real cases

    Only after you've seen the sample results, and only under a signed data agreement. The engine reads cases and prepares them; your staff decide. Access is limited to the people doing the work, and every step is logged.

  5. Running day to day

    Your systems stay the record. The engine works alongside them with the least access it needs, and your team can see what it checked, what it found and who decided, for every case.

What you can hold us to.

  • A written agreement first

    Before any real data is shared: what we receive, why, who can see it, where it's processed, how long we keep it and how it's deleted. Signed by both of us.

  • Every service named

    The agreement lists every tool and service that will touch your data, where it processes it, and its retention and training terms. If a service would use your data to train its models, we don't use it.

  • Only what's needed

    We ask for the fields a check actually uses. No bank details or client contact details unless the work genuinely requires them, and never by default.

  • People decide

    The engine never makes a clinical, legal, payment or claim decision. It prepares the case and your qualified staff decide.

  • Deleted when we're done

    At the end of a sample review, a pilot or an engagement, we delete what we hold and confirm it in writing.

  • Told straight away

    If something goes wrong with your data, we tell you as soon as we know, with the facts you need to meet your own obligations, including under the Notifiable Data Breaches scheme.

What we never do.

  • Sell or share your data.
  • Use your data to train anyone's AI models.
  • Contact your clients, participants, patients or providers.
  • Make a decision that belongs to your staff.
  • Keep data after the agreed deletion date.

Questions to ask any AI vendor.

Including us. If a vendor can't answer these plainly, that tells you something.

  • Which services will process our data, and where?
  • Do any of them keep our data, or use it to train models? For how long?
  • Who on your side can see our data, and how is that access logged?
  • Can we start with de-identified or made-up data?
  • What decisions will the system make on its own, if any?
  • How do we see what it checked and why, case by case?
  • What happens to our data when the pilot or contract ends, and how will you prove it's gone?
  • How and when will you tell us if something goes wrong?

The Office of the Australian Information Commissioner has guidance on privacy and commercially available AI products that's worth reading alongside this list.

Questions about data.

Our privacy policy covers this website and enquiries. This page covers your clients' data when we work together.

☛Do you need our client data to get started?+

No. The free session uses no data at all, and the first look at your work uses a de-identified sample. Real data only comes into it for a pilot, under a signed agreement.

☛Where is our data processed?+

That depends on the services a particular piece of work needs, so we name each one and where it processes data in your agreement before anything is shared. If a location doesn't suit you, we choose a different service or approach.

☛Is our data used to train AI?+

No. We only use services whose terms don't train on the data we send them, and the agreement names each one and its terms.

☛Who is responsible for the data?+

You stay responsible for your clients' information and your decisions. We act on your behalf, only for the work in the agreement, and we set that out in writing.

Find out how much of your caseload could run itself.

Start with a free 30-minute session. We'll learn how your cases arrive and tell you honestly how much could run on its own, and what that could mean for your growth.