Spine Master — MaxLabs.ai, MaxEnterprise, MaxMen & MaxMarketing

Start a Project

How Spine Master built Canada's first
AI-enabled physiotherapy and chiropractic platform across 6 clinic locations

Spine Master is a Canadian chain of physiotherapy, electrotherapy, and chiropractic clinics — serving patients across Ontario through in-person and hybrid care models. Recognizing that no incumbent in the Canadian market offered an integrated AI platform covering all three service lines with a shared patient record, Spine Master engaged MaxLabs to build the AI backbone, clinical intelligence features, and growth infrastructure for a multi-tenant SaaS platform: one system, six clinics, a single patient record, and AI features that extend — not replace — clinician judgment.

Spine Master — Canada's First AI Physiotherapy & Chiropractic Platform

The Challenge

Spine Master's clinicians were documenting on paper and transferring notes to an EHR manually at day's end. Physiotherapy, electrotherapy, and chiropractic ran on disconnected workflows with no shared patient record. Patient adherence to home exercise was unmonitored. No platform in the Canadian market served all three service lines with clinical AI. The chain was scaling to 10 locations — and needed an AI platform that could grow with it without compromising clinical safety, data residency, or compliance posture.

DELIVERABLES

INDUSTRY

Spine Master

Clinical AI SaaS · Physiotherapy · Canada

-50%

Doc Time Saved

≤5 days

Clinic Onboarding

AI Safety Incidents

Before — Pain Points

The key bottlenecks MaxLabs was engaged to solve:

Clinicians were documenting 100% manually. Seeing 18–22 patients per day, clinicians were writing paper notes and transferring them to the EHR at day's end. An AI ambient SOAP scribe — one that listens to the consultation, generates a structured draft note, and requires only clinician review and sign-off — was the highest-leverage single feature in the platform. Industry parallels: Penn Medicine's DAX evaluation showed 20.4% note-time reduction; Abridge across 7,260 physicians saved 15,700+ hours.
No shared patient record across service lines. A patient attending physiotherapy and chiropractic at the same clinic chain had two separate records, no shared pain history, and no cross-modality visibility for clinicians. The platform's unified data model — a single multi-tenant Postgres instance with row-level security — closes this entirely.
Home exercise adherence was unmonitored. Prescribed exercise programmes were printed or verbally communicated with no mechanism to track whether patients completed them. Without an on-device pose feedback system and adherence logging, clinicians had no signal between visits — and no basis for adjusting treatment plans between appointments.
Electrotherapy and chiropractic had no AI decision support. No competitor in the Canadian market offered electrotherapy AI protocol cross-check or structured chiropractic technique logging with outcome curves. Spine Master's platform is the first in Canada to address both service lines with clinical AI decision support, closing a category-wide gap.
Chain scaling required multi-tenant architecture from day one. Operating 6 clinics and planning 4 more, Spine Master could not run with data siloed per clinic, per-clinic billing systems, or any vendor storing PHI outside Canada. Multi-tenant design, Canadian data residency (AWS ca-central-1), and HINP-readiness were non-negotiable requirements from the first line of code.
No clinical outcome instrumentation meant no research publication path. Without structured outcome scores at every encounter, Spine Master had no data asset for internal case-series research — the differentiator that gives clinically credible platforms a moat competitors can't manufacture. The platform instruments every encounter to build the research publication pipeline alongside the SaaS business.

Our Solution

MAXLABS.AI · AMBIENT SOAP SCRIBE

Ambient SOAP scribe pipeline at the platform's core: medical ASR with diarization, Claude Sonnet 4.5 with structured-output JSON schema generating the SOAP draft, with every clinical claim evidence-span-tagged back to the encounter transcript. The clinician reviews and signs — the AI never signs. Per-patient explicit consent, audio retention capped at 24 hours, compliant with Ontario IPC AI-scribe guidance, CPSO, and CMPA requirements. Target: 50% reduction in clinician documentation time vs. paper baseline.

MAXLABS.AI · POSE & VISION ENGINE

Shared on-device vision engine powering three clinical features: posture and ROM scan at intake (phone camera, no external hardware), on-device exercise form feedback with per-repetition scoring logged to the clinician's adherence dashboard, and spinal alignment and gait analysis from smartphone video for chiropractic assessment. All pose computation runs on-device — no PHI leaves the device for vision processing. Target: 80% posture/ROM completion rate; ≥15 percentage point adherence improvement for patients with form feedback.

MAXLABS.AI · CLINICAL TRIAGE ASSISTANT

Phase 1 ships a deterministic rules-based triage assistant that handles patient symptom queries and red-flag escalation with zero LLM involvement and zero hallucination risk. Phase 2 upgrades to a RAG-powered conversational assistant grounded on the tenant's scoped clinical corpus with escalation-first guardrails. Triage red-flag precision target ≥90%, recall ≥99%. Handles patient education queries, appointment intent, and post-visit exercise FAQs — not clinical diagnosis.

MAXLABS.AI · ELECTROTHERAPY DECISION SUPPORT

Electrotherapy clinical decision support layer — a guideline-based preset selector that cross-checks the clinician's proposed protocol against evidence-based parameters, flags contraindications from intake questionnaire responses, and recommends evidence-aligned current parameters. Explicitly non-autonomous: surfaces the recommendation with evidence references; clinician accepts, modifies, or overrides. Targeting Class II SaMD submission for the adaptive recommender module in Phase 3.

MAXENTERPRISE · MULTI-TENANT SAAS INFRASTRUCTURE

Pool-model multi-tenant Postgres (Aurora) cluster in AWS ca-central-1 with row-level security enforcing per-tenant data isolation, per-tenant KMS data keys for free-text PHI, TLS 1.3 in transit, and an immutable per-event audit log with 10-year retention and nightly S3 Object Lock export. Every clinic is a tenant — same codebase, isolated data, zero cross-tenant leakage. HINP-readiness pack built before the second Ontario tenant onboarded. DR in ca-west-1. No PHI at rest or in transit outside Canada.

MAXENTERPRISE · COMPLIANCE PROGRAM

Canadian compliance program covering PIPEDA service-provider posture (clinics remain PHI custodians), provincial PHI Act compliance across PHIPA (ON), HIA (AB), HIPA (SK), PHIA (MB/NB/NS/NL/PE), and BC PIPA; HINP designation readiness; SOC 2 Type II observation beginning Phase 2 with report issued by month 17; and the Health Canada QMS (ISO 13485) and Class II SaMD submission pathway for the adaptive electrotherapy recommender and imaging pre-read module in Phase 3. Clinical safety incident target from AI features: zero.

MAXMEN · CLINICIAN COPILOT & ADMIN AUTOMATION

Operational AI layer eliminating the admin overhead that fragments clinical time: the clinician copilot surfaces patient record, prior SOAP history, and outcome trends before each encounter without navigation; the front desk booking agent handles rescheduling, intake form chasing, package payment, and reminder sequences; and the patient success seat manages adherence follow-up and re-engagement outreach for at-risk patients (target: 50%+ re-engagement rate). New clinic tenant onboarding: contract-signed to first live encounter in ≤5 business days.

MAXMARKETING · PATIENT ACQUISITION & TENANT PIPELINE

Two-track growth system: patient-facing SEO and digital acquisition targeting physiotherapy and chiropractic search demand across Ontario, and a B2B new-tenant acquisition pipeline targeting independent clinic operators considering joining the chain or licensing the platform. Phase 2's public demo surfaces — live posture scan and 3D anatomy explainer accessible without sign-up — close the gap every incumbent hides behind a registration wall. Public demos convert visitor intent into booked consultations before a prospect ever logs in.

Impact at a Glance

Clinician documentation time per session vs. paper baseline (target) −50%

Patient adherence to prescribed home exercise — rolling 30-day (target) ≥65%

Encounters with structured outcome scores — PROM/NPRS/ROM (target) ≥90%

AI scribe accuracy — notes signed with zero edits by month 12 (target) ≥75%

New clinic tenant onboarding — contract to first live encounter (target) ≤5 days

Clinical safety incidents from AI features — the only number that can't move 0

Project Results

Before MaxLabs, no platform in the Canadian market served physiotherapy, electrotherapy, and chiropractic on a shared patient record with clinical AI features. Spine Master's clinicians were documenting manually. Home exercise adherence was unmonitored. Cross-service visibility for patients didn't exist. Scaling to 10 locations meant the infrastructure question had to be answered before the clinical question could be asked.

MaxLabs built the answer across 18 months and three phases. The ambient SOAP scribe targets a 50% reduction in clinician documentation time — hours returned to patient care, not paperwork. The on-device pose and ROM engine powers three clinical features from a single vision pipeline, with no PHI leaving the device. The deterministic triage assistant handles red-flag escalation with zero LLM hallucination risk in Phase 1, upgrading to RAG in Phase 2. The electrotherapy decision support layer is the first in the Canadian market.

MaxEnterprise's multi-tenant infrastructure puts every clinic on the same codebase with isolated data, Canadian data residency, and HINP-readiness before the second tenant onboarded. The compliance program covers every provincial PHI Act, SOC 2 Type II on track for month 17, and the Health Canada SaMD pathway for Phase 3. MaxMen's admin automation targets ≤5-day new clinic onboarding and 50%+ at-risk patient re-engagement. MaxMarketing's public demo surfaces close the trust gap without a sign-up wall. The platform is the competitive moat. Every clinic added makes it stronger.

Project

Next Project

View Project

Jahangir Sons
— Retail AI

DESCRIPTION

+47% Daraz conversion, 3.1× new discovery, 62% fewer unanswered enquiries, 9× content — omnichannel AI for a Gulberg family retail store.