Sports & Fitness AI agent · Injury triage
Injury-Triage Assistant AI Agent
Take what a member describes, run the club's red-flag rules over it, and route them to a named clinician on shift — the agent records the description and the destination, and decides neither.
- Typical delivery
- 4–6 weeks
- Deployment
- Your stack
- Clinician read
- Routing only
- After launch
- Agent Care
What this agent does
Routes the member, does not assess the injury
In
- 01
Intake in the member's own words, from the club app, the front-desk kiosk, a coach or the clinic's form.
- 02
Normalisation of site, side, onset and mechanism into the fields the club's pathway expects.
Reason
- 03
Red-flag rules first, ahead of the model, on the terms the club's clinician signed.
- 04
Routing rules configured per site — who covers what, on which days, from when.
- 05
Destination checks against the current directory, so nothing routes to a service that closed.
Decide
- 06
Anything outside the configured pathways is marked unrouted, not guessed.
- 07
Handover to the named clinician or duty first-aider on shift.
Out
- 08
Retention of the description, the rule that fired, the destination and the clinician on it.
- 09
Execute write actions only inside the approval boundaries agreed during implementation.
Product statement
The agent records what was described and where it was sent; a named clinician decides what is wrong, and the club answers for the care.
Example workflow
One description, intake to pick-up
- 1Description receivedClub app, front-desk kiosk, coach referral or clinic intake form
- 2Words capturedSite, side, onset and mechanism, each kept in the words the member used for it
- 3Route draftedDestination, the rule that fired and match confidence
- 4Controls appliedRed-flag rules, pathway and staffing checks, directory verification and the confidence threshold
No human action required
Stages 1 to 4 run unaided, and nothing clinical is decided at any of them — the agent describes and routes, and the clinician's lane opens at the gate.
Goes to the named clinician on shift.
Goes to the duty clinician at once.
Clinician pick-up
The routing arrives with the description, the rule that fired and the destination.
Accept · Redirect · Escalate to urgent
Accepted — a clinician owns it
- 6Member record updatedOnly where write access and routing policy allow it
- 7Outcome evaluatedRedirect rate, time to pick-up, under-triage cases and what the clinician found
Redirects
A clinician redirect is counted in the evaluation.
What should not run autonomously
Human approval stays in control
Outside the boundary — human approval required
8 items
- Deciding what the problem is, or how serious it is.
- Telling a member to rest, train on or self-manage.
- Resolving a red-flag presentation in the conversation.
- Clearing anyone to return to play or to train.
Automation boundary
Agent acts unaided
- ✓Record what the member describes, in the words they described it in.
- ✓Run the club's red-flag rules before anything else runs.
- ✓Check the destination is one the club currently staffs.
- ✓Hand it to the named clinician on shift, and log who picked it up.
Any write happens inside the boundaries agreed at implementation, never ahead of a clinician.
- Judging whether a concussion has happened.
- Naming a condition, a tissue or a likely cause.
- Interpreting a scan, a wearable stream or a test.
- Changes to the red-flag list or the pathways.
Example output
One description, annotated
Everything the agent sends on is attached to the words it was drawn from.
Routing output · single description
Illustrative example- Member
- Adult, gym floor
- Described as
- Rolled my ankle on a box jump, it is swelling, I can stand on it
- Reported at
- 18:42 local
- Routed to
- Club physio clinic
- Confidence
- 88%
- Rule fired
- No red-flag rule fired
As received
Taken from the member's own words and the time it was entered — nothing on this side is interpreted.
What was captured
- Site, side and onset
- Mechanism given
- Time entered
Why this route
It matches the pathway the club configured for that site — and no cause is named.
What the score decides
Below the configured threshold, or on any red-flag rule, it goes to the duty clinician at once.
Action
AcceptRedirectEscalate to urgentValue
Where AI adds value
The same four claims, placed at the point in the workflow where each one applies.
Where the value lands
Value 01 – 04
Every description
From the member's words
Match the pathway
Work from the red-flag rules the club's clinician signed and the pathways it staffs.
Route it, do not read it
The description reaches a named clinician without a queue to watch.
Send red flags straight through
Red-flag rules fire ahead of the model, so a presentation that cannot wait moves before anything is weighed.
Build an evidence trail
What was described, what routed it and who picked it up stay on the record.
Integrations
Typical integrations
Five system groups connect to the same agent. Which of them are in scope is decided in discovery.
Club management
- Mindbody
- Glofox
- ABC Fitness
- Perfect Gym
Member apps
- Club app
- web intake form
- Front-desk kiosk
- staff entry
Clinic systems
- Cliniko
- Nookal
- Halaxy
- Jane
- TM3
- practice diary
Agent
Injury-triage routing
- Takes the words
- Runs the rules
- Routes to a person
Athlete records
- Kitman
- Smartabase
- Teamworks
- Edge10
Observability & evaluation
- OpenTelemetry
- Langfuse
- Supported monitoring/evaluation sources
Integration availability depends on the client's existing systems and API access.
Agent controls
Six layers between the model and the member
Every layer wraps the next. What none of them catches is in the map below.
L6 · Outermost — last line of defence
Inward → L1 · closest to the model
- L1 – L2
- Test whether a routing stands
- L3
- Puts the description in front of a person
- L4 – L5
- Hold the description the routing rested on
- L6
- Falls back to book-a-clinician when signals slip
How Nestack evaluates it
Evaluate the routing workflow — not only the destination it picked.
Coverage runs the whole depth of the workflow, and every layer is cut by slice.
Surface — what the member is told
- E1Final-output evaluationDid the routing match the pathway the club configured?
- E2Step-level evaluationDid the agent use the description given, the current rules and the live directory?
- E3Tool evaluationDid it write to the right member record and the right clinic diary?
- E4Confidence calibrationDo low-confidence routings actually attract more redirects?
- E5Slice evaluationHow does performance change across presentation types?
- E6Business outcomeHow many routings were redirected, and how many were under-triaged?
Floor — the outcome the club answers for
Failure modes
Where each failure originates in the agent
Seven failure modes, each at the stage where it starts.
Agent lifecycle
Direction of processing →
01 · Retrieval
1 mode
- JT-03
Stale pathway entry
Routed to a clinic the club no longer staffs.
Stage gathers
The description, the red-flag list and the directory
02 · Reasoning
2 modes
- JT-04
Under-triage
A red-flag description is routed as a routine one.
- JT-06
Named cause
The reply names a condition instead of a destination.
Stage proposes
A destination, the rule that fired and confidence
03 · Tool / write
2 modes
- JT-02
Routed without a person
Agent completes a routing that needed a clinician.
- JT-05
Duplicate intake
One description is routed twice, to two destinations.
Stage writes
Only where write access and approval policy allow it
04 · Output
1 mode
- JT-01
Wrong destination
The member is sent to a pathway that cannot help.
Stage returns
The destination the member is given and acts on
05 · Change / Version
1 mode
- JT-07
Silent rule regression
A model or rule change softens what counts as a red flag.
Stage tracks
Model, prompt, red-flag rules and directory changes
- Sev-1 · the agent acts outside the boundary
- Sev-2 · a red-flag description routes as routine
- Sev-3 · input degrades, intake goes to a person
Affected slices
The aggregate buries one presentation
Aggregate redirect rates hold steady while back pain with leg symptoms absorbs the redirects. Nestack reports the clinician-redirect rate by presentation type, not only in total.
Slice performance — reported separately, not only in aggregate
Illustrative example| Slice | Failure rate | Lift | Lift vs. threshold | Status |
|---|---|---|---|---|
| Back pain with leg symptoms | 7.8% | 3.8× | Review | |
| Head or neck knocks | 5.9% | 2.9× | Review | |
| Returning after a layoff | 4.1% | 2.0× | Watch | |
| Simple strains and knocks | 1.8% | 0.9× | Normal |
Bar: clinician-redirect-rate lift vs. simple-strain baseline · scale 0–4.0× · tick marks the 2.0× review threshold
2 of 4 slices over threshold
Evidence-linked improvement
What happens after a bad routing
The loop shuts when the miss is a case in the suite, not when it has been explained. That suite is what the next routing out of intake is measured against.
Improvement cycle · five stages
Switchback — the path turns at Improve and returns at Learn
- 01
Detect
Redirect rate rises in a presentation type.
- 02
Diagnose
Which rule fired, and on what words? A clinician reads back until one cause is left.
- 03
Improve
Whatever changes ships against a version, with the presentations that prompted it attached.
- 04
Verify
Nothing ships until the affected cases pass a second time.
- 05
Learn
The suite grows by one case; so does the red-flag list.
Learn → Detect
The return edge. What the next miss is measured against has grown by one.
Typical build scope
Twelve workstreams across six weeks
The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, sources, routing workflow, evaluation, integration, then production validation and handover.
Workstream
Week 1Week 2Week 3Week 4Week 5Week 6
- 01Intake workflow discovery and boundary definition.Week 1
- 02Club, app and clinic source assessment.Week 1, Week 2
- 03Red-flag rules and referral-pathway directory mapping.Week 1, Week 2
- 04Intake ingestion and description normalisation.Week 2
- 05Routing logic and red-flag precedence.Week 2, Week 3
- 06Confidence scoring and duty-clinician routing.Week 3
- 07Clinician pick-up workflow.Week 3, Week 4
- 08Clinic and club-system integration.Week 5
- 09Under-triage regression cases.Week 4, Week 5
- 10Guardrails and red-flag controls.Week 4
- 11Routing-trail instrumentation.Week 5, Week 6
- 12Deployment, documentation and Agent Care handover.Week 6
12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallel
Final scope and sequence confirmed in discovery
Engagement tiers
What each tier includes
Rows are the capabilities named in each tier's scope. Higher tiers include everything below them.
| Capability✓ in scope · — not at this tier | PilotOne site, one clinic | ProductionProduction club systems | AdvancedMultiple sites / brands |
|---|---|---|---|
| Introduced at Pilot | |||
| Routing to your rules and pathways | ✓ | ✓ | ✓ |
| Clinician pick-up | ✓ | ✓ | ✓ |
| Routing-consistency baseline | ✓ | ✓ | ✓ |
| Introduced at Production | |||
| Reporting by presentation type | — | ✓ | ✓ |
| Pick-up workflow in your systems | — | ✓ | ✓ |
| Approved write-back | — | ✓ | ✓ |
| Practice-management integration | — | ✓ | ✓ |
| Introduced at Advanced | |||
| Multi-site pathway rules | — | — | ✓ |
| Cross-site clinician rotas | — | — | ✓ |
| High intake volume | — | — | ✓ |
| Multi-clinic routing controls | — | — | ✓ |
| Build price | From $5,000 | From $8,000 | Custom quote |
- Final build price
- Confirmed after discovery based on integrations, workflow complexity, transaction volume, approval controls and deployment requirements.
- Separate from build
- Build pricing is separate from recurring Agent Care, which covers managed monitoring, evaluations, incidents and verified improvements after launch.
What we need from you
What you bring, and what we build with it
Each input maps to a piece of build scope and a week in the delivery timeline.
You bring
We build with it
01Your intake fields and description structure
→Intake ingestion and description mapping
Week 102Representative member intakes
→Routing baseline, rule extraction and destination binding
Week 203Your red-flag list, signed by a clinician
→Red-flag rules and referral-pathway directory mapping
Week 104Access to relevant APIs, feeds or exports
→Club, app and clinic assessment, then integration setup
Week 205Routings that sent someone the wrong way
→Under-triage cases and failure-mode testing
Week 406Where description stops and assessment begins
→Confidence scoring, duty-clinician routing, guardrails and controls
Week 307Named clinicians to pick intakes up
→Clinician pick-up workflow, then pilot and production validation
Weeks 5–6
Nothing else is required
Deployment, documentation and Agent Care handover are ours.
Delivery timeline
Four phases across six weeks
The bands follow real work rather than a plan, so evaluation and pilot genuinely share the fifth week.
Phase
W1
W2
W3
W4
W5
W6
- DiscoveryW1
- BuildW2 – W3
- EvaluateW4 – W5
- Pilot & LaunchW5 – W6
Week focus
W1Intake workflow discovery, red-flag mapping and the automation boundary
W2Source integration and the routing baseline
W3Routing workflow, confidence logic and pick-up controls
W4Evaluation suite, red-flag checks and under-triage failure testing
W5Clinic integration, pilot intakes and targeted corrections
W6A live intake queue routed under clinician supervision, then handover
- Reading the band
- Each bar covers only the weeks its work is named in. The fifth week genuinely carries two kinds of work.
- At the end of W6
- Once the queue validates, Agent Care owns the running agent.
- Duration
- Six-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.
Next step · Sports & Fitness agent
Build an injury-routing agent around the clinicians you already have.
Show us how a member reports an injury today and who picks it up. You bring the red-flag list a clinician has signed, the pathways you actually staff and the names on shift.
- Typical delivery
- 4–6 weeks
- Deployment
- Your club systems
- Clinician read
- Routing only
- After launch
- Agent Care available
Nestack Agents · Injury-triage routing
AGT-SF-03 · Agent Care available after launch