← All work

Healthcare·2026

Veterinary Journey Audit

An as-is audit of our own veterinary module, tracing one patient from the reception desk to the final payment to find where the clinical work and the money stop talking to each other.

Client
NewBranch (own veterinary module)
Role
Research, service and interface audit
Sector
Healthcare
Status
Audit complete, August 2026
A veterinarian examining a dog while a colleague works at a computer

The challenge

Before automating anything in a clinic, you need to know how the work really moves between people and screens. A feature that exists in the database is not the same as a step a receptionist can reach at eight in the morning.

Our approach

We walked the full journey in a development build with a seeded demonstration practice, logged every screen against what should have come next, and recorded each break as a gap with its cause. Nothing was redesigned until the evidence was written down.

In brief

12

findings, two of them things that already work well

7

critical gaps, each tagged as an interface or an integration fix

1 of 8

hand-offs between stages judged sound

What we audited

Our own veterinary module, from the front desk to the final payment.

NewBranch OS has a veterinary vertical that adds clinical records to the shared platform. An earlier read-only review had found that the vertical held scheduling settings and vocabulary but no clinical storage. Once the clinical workspace was built, we audited the journey a practice would actually live through.

The audit ran in August 2026 on a development build, against a seeded demonstration practice with one location. We followed a single patient in a chronic-care episode, with one owner and one staff login acting as both receptionist and clinician. No real clinic, patient or owner data was involved. Nine stages were in scope.

  1. 01Reception boardObserved
  2. 02Intake routesOpened, form not submitted
  3. 03Patient workspaceObserved
  4. 04Clinical encounterObserved on a seeded encounter
  5. 05DocumentationFrom the earlier review
  6. 06CompletionNot shown
  7. 07Medicine consequenceNot shown
  8. 08Billing hand-offNo path from clinical screens
  9. 09PaymentNo paid invoices to trace
The nine stages in scope, and how far the walkthrough could observe each one.

How we worked

Evidence first, opinions second.

Every screen was logged with the person who uses it, the screen that should come next, the screen that actually came next and an assessment. Each gap records two things separately: whether the underlying capability exists, and whether any screen in the product reaches it. That split tells the team whether the fix is interface work or integration work.

Where a stage could not be observed in the walkthrough, it was marked as not captured and moved to the gap register, rather than described from what the code suggested. The audit documents what the product delivers today, not what it is capable of in principle.

What we found

Strong screens that do not hand off to each other.

The product behaved like several modules that happen to hold related data, not one continuous veterinary workflow. Some parts were good. Patient context held throughout, and an animal is never filed under its owner's identity. The encounter workspace, the reception board, the patient record and the signing of triage notes all work. The patient record also says plainly what it does not know: it flagged that half of the context facts, including weight and consent authority, could not be established.

The most important finding was that clinical work produces no visible financial consequence. There is no running total of charges, no action to generate an invoice and no signal to reception that a patient is ready to check out. Billing lives in a separate finance module. Of the eight hand-offs we assessed, only triage to encounter was sound.

  1. 01Reception to bookingNo visible confirmation
  2. 02Booking to arrivalNo arrival step
  3. 03Arrival to triageNo capture by reception
  4. 04Triage to encounterSound
  5. 05Encounter to chargesCritical break
  6. 06Charges to invoiceModule boundary
  7. 07Invoice to paymentModule boundary
  8. 08Payment to closureNo signal back
The eight hand-offs between stages, as assessed in the audit.

Clinical risk

Where a workflow gap becomes a safety question.

Several findings matter beyond efficiency. Medicines and stock are well founded but only partly connected, so a clinician could prescribe something that is out of stock and not know it. Important clinical information sat below the fold, which risks decisions made on an incomplete picture. Responsibility between reception and clinician was unclear at the moment of hand-off, including who re-triages a patient whose condition changes while waiting.

Some screens were designed around database capabilities rather than the person using them, to the point of showing internal specification references in the interface. Booking context was lost after the first record, and location context disappeared once an encounter began.

Recommendations

Choreography, not technology.

The modules exist. What is missing is the movement between them, so the five recommendations connect existing screens rather than add new modules. They are written as the scope of a proof of concept, in this order:

  • Show the booking confirmation straight after walk-in intake
  • Make arrival a single, obvious step on the reception board
  • Give reception a clickable ready-for-checkout signal from the clinic
  • Show estimated charges during clinical work, without hiding or changing clinical data
  • Replace scattered next steps with one view of what this patient needs next
A veterinarian and assistant with a patient
A veterinarian examining a pet with a stethoscope

Key decisions

What we chose, why, and what it cost us.

  1. 01

    Audit what the product delivers today before redesigning anything.

    Why

    In a clinic, a capability buried in the database helps nobody. Redesigning from the code's point of view would have repeated the problem the audit found.

    Trade-off

    The result is an uncomfortable read about our own product. Stages without a working screen appear as gaps, not features, even where the underlying data model is in place.

  2. 02

    Trace one seeded patient end to end instead of creating fresh bookings.

    Why

    Creating new bookings adds form-filling time that says little about continuity. An existing encounter shows whether the product carries one patient's story from desk to payment.

    Trade-off

    The trace started mid-journey, so booking completion and arrival could not be shown and went to the gap register. With one login playing both roles, the hand-off between people was judged from the screens, not watched.

  3. 03

    Record whether a capability exists separately from whether anyone can reach it.

    Why

    The same gap can be a missing button or a missing integration. Splitting the two tells the build team where the work actually is.

    Trade-off

    Each gap takes longer to document, and several could only be classified as needing both interface and integration work.

  4. 04

    Use a seeded demonstration practice, not a live clinic.

    Why

    It let us run the audit on a development build with no real patient, owner or staff data involved.

    Trade-off

    The seed data held no paid invoices and there was no live banking connection, so payment and reconciliation could not be tested end to end. The billing conclusions rest on the absence of a path from the clinical screens, not on a tour of the finance module.

  5. 05

    Make charges visible in clinical screens, but never let money block care.

    Why

    The biggest break was that clinical work created no visible charge. The fix belongs where the work happens, not in a separate finance screen.

    Trade-off

    Money in a clinical view can put pressure on clinical judgement. The platform's clinical rules already state that emergency care must not be blocked by a financial flag, and the recommendation keeps clinical data untouched.

Where it stands

The audit is finished and its findings stand as the baseline for the next build phase. It covers our own product on demonstration data, so it shows how we audit, not results at a client practice.

What we would do next

  • 01

    Build the five recommendations as a proof of concept and re-run the same trace against it.

  • 02

    Complete the stages the walkthrough could not reach, starting with booking, arrival and the finance module.

  • 03

    Seed paid invoices and test payment recording and reconciliation end to end.

  • 04

    Run the journey with separate reception and clinician logins to observe the hand-off between people.

  • 05

    Repeat the audit with a working practice, with its consent, before any automation goes live.

Written from the as-is journey audit (flow map, findings and gap register, August 2026) and the earlier clinical system review. Seeded names, identifiers and internal routes are left out on purpose.