Capex vs Opex AI Decisions in Allied Health Portcos

Table of Contents

1. Why This Matters for Allied Health PE
2. The CapEx vs OpEx Decision Framework
3. AI Capabilities Map: What to Buy, Build, or Rent
4. Diligence Playbook: Assessing Current AI Posture
5. Value Creation Levers Through AI
6. Building Your 100-Day AI Roadmap
7. Compliance, Risk, and Audit Readiness
8. Portfolio Company Benchmarks and KPIs
9. Exit Positioning: Making AI Visible to Buyers
10. Getting Started: Next Steps

---

Why This Matters for Allied Health PE

Allied health is one of the highest-growth, most fragmented sectors in healthcare. Physiotherapy, dental, pathology, aged care, psychology, and diagnostic imaging practices are being rolled up into regional and national platforms. Yet most portfolio companies are running on legacy systems, paper workflows, and siloed data.

AI is reshaping the economics of these businesses—but not in the way most founders think. It’s not about replacing clinicians. It’s about compressing admin time, reducing no-shows, automating scheduling and billing, improving diagnosis confidence, and freeing practitioners to focus on higher-value patient interaction.

The problem: most allied health operators face a false choice between buying expensive, rigid SaaS platforms (CapEx-heavy, long implementation, limited to pre-built workflows) and building custom AI from scratch (expensive engineers, long timelines, ongoing maintenance burden).

This guide gives you a third way. We’ll walk through the decision framework, diligence questions, and value-creation playbook that allows you to move fast, cut costs, and position your portfolio for exit—without betting the company on AI.

---

The CapEx vs OpEx Decision Framework

Understanding the Spectrum

CapEx and OpEx aren’t binary. They sit on a spectrum, and the best AI decisions often blend both.

CapEx-Heavy Approach:

• Buy or build proprietary AI software (e.g., custom diagnostic model, proprietary scheduling system)
• Invest in infrastructure (servers, GPUs, data warehouses)
• Upfront cost: AU$500K–AU$5M+
• Payback: 2–4 years (if it works)
• Risk: High. If the market changes or adoption stalls, you’re stuck with stranded assets
• Ownership: You own the IP and the technical debt

OpEx-Heavy Approach:

• Use SaaS AI tools (ChatGPT, Claude, specialist health AI platforms)
• Outsource implementation to a vendor or agency
• Monthly/annual cost: AU$5K–AU$50K+ per application
• Payback: Immediate (30–90 days)
• Risk: Medium. Vendor lock-in, feature limitations, data residency concerns
• Ownership: Vendor owns the model; you own the workflow

Hybrid (Recommended for Most Allied Health Portcos):

• Use best-of-breed SaaS AI for commodity tasks (scheduling, billing, document processing)
• Build or co-build custom AI only for defensible, high-value workflows (clinical decision support, patient outcome tracking)
• Fractional CTO or agency partnership for architecture and delivery
• Upfront: AU$50K–AU$300K
• Monthly: AU$10K–AU$30K
• Payback: 6–18 months
• Risk: Low-to-medium. Modular, testable, can pivot
• Ownership: You own the integration and the clinical workflows; SaaS vendors own the commodity layers

The Decision Rubric

For each AI use case in your portfolio, ask:

1. Is this a competitive advantage? (Does it differentiate you from competitors, improve patient outcomes, or unlock new revenue?)

   • Yes → Consider CapEx or hybrid build
   • No → Use OpEx SaaS

2. Is the workflow stable or changing? (Will the rules stay the same for 2+ years?)

   • Stable → CapEx or hybrid build (amortise investment)
   • Changing → OpEx SaaS (flexibility)

3. How much data do you have? (Can you train or fine-tune a custom model?)

   • 10K+ labelled examples → Consider custom build
   • <5K examples → Use pre-trained SaaS models

4. What’s the cost of failure? (If the AI gets it wrong, what happens?)

   • High (clinical decision support) → Hybrid or OpEx with human oversight
   • Low (scheduling suggestion) → OpEx SaaS alone

5. Do you have engineering capacity? (Can you hire, onboard, and retain engineers?)

   • Yes → Hybrid or CapEx
   • No → OpEx + fractional CTO or agency partner

Real Allied Health Example: Scheduling + Clinical Notes

Scheduling (low competitive advantage, stable workflow, low cost of failure):

• OpEx play: Use a SaaS scheduling tool with AI no-show prediction (e.g., Calendly + custom Zapier logic, or a health-specific platform like Acuity or Mindbody)
• Cost: AU$200–500/month
• ROI: 30–50% reduction in no-shows = 5–10% revenue uplift
• Payback: 1–2 months

Clinical notes (high competitive advantage, stable workflow, high cost of failure):

• Hybrid play: Use a SaaS transcription + templating tool (e.g., Otter Health, Ambience, or Microsoft Teams transcription), then build a custom clinical note summariser using an open-source LLM (e.g., Llama 2) or a fine-tuned GPT model
• Cost: AU$500/month SaaS + AU$100K one-time build + AU$5K/month operations
• ROI: 20–30% time saving for clinicians = 15–20% productivity uplift
• Payback: 8–12 months

The hybrid approach lets you move fast (use SaaS transcription immediately), reduce risk (test clinical summarisation with a small cohort first), and build defensibility (your note templates and clinical logic are proprietary).

---

AI Capabilities Map: What to Buy, Build, or Rent

Here’s a practical map for the most common allied health AI use cases. We’ve scored each on strategic value, implementation complexity, and recommended approach.

Patient-Facing AI (High Strategic Value)

Appointment Scheduling & No-Show Prediction

• Strategic value: Medium (revenue protection)
• Complexity: Low
• Recommendation: OpEx SaaS + integration
• Typical cost: AU$300–1K/month
• Timeline: 2–4 weeks
• Example vendors: Calendly, Mindbody, Acuity, or custom Zapier workflows

Patient Intake & Triage

• Strategic value: High (reduces admin, improves triage accuracy)
• Complexity: Medium (requires clinical validation)
• Recommendation: Hybrid (SaaS form builder + custom logic)
• Typical cost: AU$1K–5K/month SaaS + AU$30K–100K build
• Timeline: 4–8 weeks
• Example: Use Typeform or Jotform for intake, then build custom triage logic using Claude API or GPT-4 to flag high-risk cases

Telehealth/Video Consultation AI

• Strategic value: High (unlocks remote revenue, improves access)
• Complexity: Medium-to-high (HIPAA/privacy, latency)
• Recommendation: OpEx SaaS (don’t build video infrastructure)
• Typical cost: AU$2K–10K/month
• Timeline: 2–4 weeks
• Example vendors: Teladoc, MDLive, or white-label platforms like Amwell or TidalHealth

Clinical AI (Highest Strategic Value, Highest Risk)

Diagnostic Decision Support (Imaging, Pathology)

• Strategic value: Very high (clinical outcomes, liability reduction)
• Complexity: Very high (regulatory, data quality, clinical validation)
• Recommendation: Hybrid or CapEx with regulatory expertise
• Typical cost: AU$200K–1M+ for build + ongoing compliance
• Timeline: 3–6 months (including validation)
• Approach: Partner with a vendor like Zebra Medical Vision or use a research model as a starting point, then build clinical validation framework
• Note: This is where you must engage a Fractional CTO & CTO Advisory in Boston or equivalent if you’re in a regulated market, or a Fractional CTO & CTO Advisory in Sydney if you’re Australia-based. Clinical AI is not a DIY project.

Patient Outcome Tracking & Prediction

• Strategic value: High (improves care, enables value-based contracting)
• Complexity: High (data quality, clinical validation)
• Recommendation: Hybrid (use SaaS EHR/EMR data, build custom prediction model)
• Typical cost: AU$50K–200K build + AU$5K–15K/month
• Timeline: 8–12 weeks
• Example: Extract patient data from your EHR (e.g., Healthengine, Argus, or Zedmed), build a custom outcome prediction model using Python/ML stack, deploy as a dashboard for clinicians

Operational AI (High ROI, Lower Risk)

Billing & Revenue Cycle Automation

• Strategic value: High (cash flow, reduced write-offs)
• Complexity: Medium (data integration, claim rules)
• Recommendation: OpEx SaaS or hybrid
• Typical cost: AU$1K–5K/month SaaS
• Timeline: 2–6 weeks
• Example vendors: Athenahealth, Medidata, or custom RPA + AI (e.g., UiPath + GPT-4 for claim denial analysis)

Staff Scheduling & Rostering

• Strategic value: Medium (labour cost, retention)
• Complexity: Low-to-medium (rules-based, but many edge cases)
• Recommendation: OpEx SaaS or lightweight custom build
• Typical cost: AU$500–2K/month SaaS
• Timeline: 2–4 weeks
• Example vendors: Deputy, Zip Schedules, or custom Zapier workflows

Document Processing & Data Entry Automation

• Strategic value: Medium (admin cost reduction)
• Complexity: Low (RPA + OCR + LLM)
• Recommendation: OpEx SaaS or lightweight custom build
• Typical cost: AU$500–2K/month SaaS
• Timeline: 1–2 weeks
• Example: Use Zapier + ChatGPT API to extract data from referral letters, consent forms, or medical records

Email & Communication Triage

• Strategic value: Low-to-medium (admin efficiency)
• Complexity: Low
• Recommendation: OpEx SaaS
• Typical cost: AU$200–500/month
• Timeline: 1 week
• Example vendors: Gmail + SaneBox, or Slack + Slackbot

---

Diligence Playbook: Assessing Current AI Posture

When you’re evaluating an allied health acquisition, AI readiness should be part of your tech diligence. Here’s what to look for.

The 90-Minute Diagnostic

Before you engage engineers, spend 90 minutes with the CEO, CTO, and head of operations. Ask:

1. What AI or automation are you already using?

   • Look for: Off-the-shelf SaaS (good sign of tech-friendliness), custom builds (good sign of ambition, but check execution), or nothing (opportunity)
   • Red flag: “We don’t do tech” or “Our EHR vendor handles everything”

2. What’s your biggest operational pain point?

   • Listen for: No-shows, billing delays, admin overhead, clinical inefficiency
   • This is your first value-creation lever

3. Do you have data?

   • Ask: “Can you export your last 12 months of patient records, scheduling, and billing data?”
   • If yes, you have optionality. If no, you’ll need to build data infrastructure first.

4. Who’s in charge of tech decisions?

   • If it’s the founder/CEO: Fast decision-making, but may lack technical depth
   • If it’s a CTO/IT manager: Slower decisions, but more technical credibility
   • If there’s no one: Red flag. You’ll need to hire or partner.

5. What’s your EHR/practice management system?

   • Modern systems (Healthengine, Argus, Zedmed, Helix, Telstra Health): Good API support, easier to integrate AI
   • Legacy systems (paper, old desktop software): High integration cost, but also high value creation

The Technical Audit

If the deal is serious, do a deeper technical audit. Allocate 3–5 days and AU$10K–20K.

Data Audit:

• How much patient data do you have? (Aim for 2+ years, 1K+ records)
• What’s the quality? (Missing fields, inconsistent formats, duplicates?)
• Where does it live? (EHR, spreadsheets, paper?)
• Can you export it? (API, CSV, or manual process?)
• Is it HIPAA/GDPR-compliant? (Encryption, access controls?)

System Audit:

• What’s your current tech stack? (EHR, billing, CRM, scheduling?)
• Are systems integrated or siloed?
• What APIs are available?
• What’s the uptime/reliability?

Compliance Audit:

• Do you have a security policy?
• Are you ISO 27001 or SOC 2 certified? (Increasingly required by large customers and insurers)
• Have you had a security audit in the last 2 years?
• Are you GDPR/HIPAA-compliant?

If compliance is weak, budget an additional AU$50K–150K for a Security Audit (SOC 2 / ISO 27001) via a partner like PADISO. This is a value-creation play: better compliance = higher exit valuation, especially if you’re selling to a larger health platform.

The AI Readiness Checklist

Score each item 0–2 (0 = no, 1 = partial, 2 = yes). Aim for 12+ to move forward with confidence.

• Clear operational pain point identified (e.g., 30% no-show rate, 20 hours/week admin)
• Data available and exportable (2+ years, 1K+ records)
• EHR/PMS is modern and API-enabled
• Executive sponsor identified (CEO or COO)
• Budget allocated (AU$50K–300K for first 12 months)
• Clinical/operational team willing to pilot AI
• Compliance baseline established (at least privacy policy)
• No major IT infrastructure debt (e.g., not running on 20-year-old servers)

---

Value Creation Levers Through AI

Here’s where the money is. We’ve worked with 50+ allied health and health-tech portfolio companies. These are the proven value levers.

Lever 1: No-Show Reduction (Quick Win, 6–12 Month Payback)

The Problem: Allied health has a 20–30% no-show rate. Each no-show is lost revenue (typically AU$80–200 per slot). A 20-bed physio clinic with 80 appointments/week loses AU$1,280–3,200/week to no-shows = AU$65K–165K/year.

The AI Solution:

• Implement predictive no-show scoring using historical booking, patient, and outcome data
• Trigger automated SMS/email reminders 24–48 hours before appointment (personalised by risk tier)
• Optionally: offer incentives (e.g., “Book your next appointment now and save AU$20”) to high-risk patients

The Economics:

• Cost: AU$300–1K/month SaaS + AU$5K–10K integration
• Expected impact: 20–40% reduction in no-shows (conservative: 25%)
• Revenue uplift: AU$16K–41K/year
• Payback: 2–4 months
• Ongoing: AU$3K–5K/year

Implementation Path:

1. Week 1: Export 12 months of booking and outcome data
2. Week 2–3: Build no-show prediction model (Python + scikit-learn or use a SaaS tool like Calendly’s built-in prediction)
3. Week 4: Integrate with your SMS/email platform (e.g., Twilio, Mailchimp)
4. Week 5+: A/B test reminder cadence and messaging

Lever 2: Clinical Productivity (12–18 Month Payback)

The Problem: Clinicians spend 30–40% of their time on admin: note-taking, documentation, care planning, referral writing. A physiotherapist billing at AU$150/hour loses AU$15K–20K/year to admin overhead.

The AI Solution:

• Implement voice-to-note transcription (e.g., Otter Health, Ambience, or Microsoft Teams transcription)
• Build custom clinical note summariser that extracts key findings, creates care plans, and flags follow-up actions
• Integrate with EHR for one-click documentation

The Economics:

• Cost: AU$500–1K/month SaaS + AU$30K–80K custom build + AU$3K–5K/month operations
• Expected impact: 20–30% reduction in admin time (conservative: 20%)
• Productivity uplift: AU$10K–15K per clinician per year
• Payback: 8–12 months (for a 10-clinician practice)
• Ongoing: AU$5K–10K/year

Implementation Path:

1. Week 1–2: Pilot transcription with 2–3 clinicians (test audio quality, privacy)
2. Week 3–4: Build note summariser using Claude API or fine-tuned GPT-4
3. Week 5–6: Integrate with EHR and create clinician workflows
4. Week 7+: Gather feedback, iterate, roll out to full team

Lever 3: Revenue Cycle Optimization (6–12 Month Payback)

The Problem: Billing errors, claim denials, and slow collections cost allied health practices 5–10% of gross revenue. A AU$2M practice loses AU$100K–200K/year.

The AI Solution:

• Implement automated claim validation (check for missing fields, coding errors, eligibility)
• Build AI-powered denial analysis (flag high-risk claims before submission)
• Automate follow-up reminders for unpaid invoices

The Economics:

• Cost: AU$1K–3K/month SaaS + AU$10K–30K integration
• Expected impact: 30–50% reduction in claim denials (conservative: 30%)
• Revenue recovery: AU$30K–100K/year
• Payback: 3–6 months
• Ongoing: AU$2K–5K/year

Implementation Path:

1. Week 1: Audit last 12 months of claims (identify top denial reasons)
2. Week 2–3: Set up automated validation rules in your billing system
3. Week 4–5: Build denial prediction model or integrate with a SaaS tool
4. Week 6+: Monitor and iterate

Lever 4: Patient Outcomes & Retention (18–24 Month Payback)

The Problem: Allied health has a 30–40% patient dropout rate. Patients often don’t complete their course of treatment, leading to poor outcomes and lost revenue.

The AI Solution:

• Build outcome prediction model to identify at-risk patients (likely to drop out)
• Trigger automated outreach (SMS, email, clinician call) to at-risk patients
• Create personalised care plans based on predicted outcomes

The Economics:

• Cost: AU$50K–150K custom build + AU$3K–5K/month operations
• Expected impact: 15–25% improvement in completion rates (conservative: 15%)
• Revenue uplift: AU$30K–100K/year (depends on practice size and pricing)
• Payback: 12–18 months
• Ongoing: AU$3K–5K/year

Implementation Path:

1. Week 1–2: Extract outcome data from EHR (discharge status, functional improvement, patient satisfaction)
2. Week 3–4: Build prediction model (identify features that predict dropout)
3. Week 5–6: Design intervention workflow (who reaches out, what do they say?)
4. Week 7–8: Pilot with high-risk cohort
5. Week 9+: Scale and measure impact

Lever 5: Scalable Service Delivery (Ongoing ROI)

The Problem: Allied health is labour-constrained. You can’t grow revenue without hiring more clinicians, and clinicians are expensive and hard to find.

The AI Solution:

• Implement telehealth/video consultation capability (extend reach beyond geography)
• Build patient self-service portal (education, exercises, symptom tracking)
• Create AI-powered patient education (personalised exercise videos, progress tracking)

The Economics:

• Cost: AU$2K–10K/month SaaS + AU$20K–50K integration
• Expected impact: 20–40% increase in patient capacity (without hiring)
• Revenue uplift: AU$100K–500K/year (depends on practice size and pricing)
• Payback: 6–12 months
• Ongoing: AU$2K–10K/year

Implementation Path:

1. Week 1: Choose telehealth platform (e.g., Zoom Health, Teladoc, or white-label)
2. Week 2–3: Set up infrastructure and train clinicians
3. Week 4+: Market to patients, monitor utilisation and outcomes

---

Building Your 100-Day AI Roadmap

You’ve identified your portfolio company. You’ve done diligence. Now it’s time to move. Here’s a battle-tested 100-day roadmap.

Days 1–14: Mobilisation & Quick Wins

Objective: Build momentum, secure internal buy-in, identify first value lever.

Activities:

• Day 1–2: Kick-off meeting with portfolio company leadership (CEO, COO, head clinician)
• Day 3–5: Data export and audit (can you get the data you need?)
• Day 6–7: Compliance baseline assessment (privacy, security, SOC 2 readiness)
• Day 8–10: Identify top 3 pain points and quick wins
• Day 11–14: Launch first quick win (e.g., no-show prediction, scheduling optimisation)

Deliverables:

• 100-day roadmap (3–5 initiatives, phased)
• Data audit report (quality, volume, integration readiness)
• Compliance gap analysis
• First quick-win project plan (timeline, budget, success metrics)

Budget: AU$5K–15K (internal team + initial data work)

Days 15–50: First Value Creation Initiative

Objective: Ship first AI initiative, prove ROI, build internal credibility.

Example: No-Show Prediction

Timeline:

• Day 15–20: Data preparation (clean, label, train/test split)
• Day 21–30: Model building (use SaaS tool or custom ML)
• Day 31–40: Integration with scheduling + SMS/email
• Day 41–50: Pilot with 20% of appointments, A/B test messaging

Success Metrics:

• Model accuracy: >75% precision on high-risk predictions
• Reminder click-through rate: >40%
• No-show rate reduction: >15% in pilot

Budget: AU$15K–30K (tools, integration, data science time)

Alternative: Clinical Productivity (Transcription + Note Summarisation)

Timeline:

• Day 15–20: Transcription tool pilot (2–3 clinicians, test audio quality)
• Day 21–35: Note summariser build (Claude API or fine-tuned GPT-4)
• Day 36–45: EHR integration and clinician training
• Day 46–50: Pilot feedback and iteration

Success Metrics:

• Transcription accuracy: >95% (medical terms)
• Clinician time savings: >20% on documentation
• Note quality (clinician satisfaction): >4/5

Budget: AU$30K–60K (SaaS tools, custom development, integration)

Days 51–80: Scale First Initiative + Launch Second

Objective: Roll out first initiative to full user base, launch second value lever.

Activities:

• Days 51–60: Monitor first initiative metrics, iterate based on feedback
• Days 61–70: Full rollout of first initiative (all users, all locations)
• Days 71–80: Launch second initiative (e.g., revenue cycle optimisation, outcome prediction)

Success Metrics:

• First initiative: Achieving target metrics across full user base
• Second initiative: Pilot launched, early feedback positive

Budget: AU$20K–40K (rollout, training, second initiative setup)

Days 81–100: Measurement, Documentation, Planning

Objective: Quantify ROI, document playbooks, plan next 12 months.

Activities:

• Days 81–90: Comprehensive impact measurement (revenue uplift, cost savings, clinician satisfaction)
• Days 91–95: Document playbooks, training materials, runbooks
• Days 96–100: Plan next 12 months (roadmap, budget, hiring)

Deliverables:

• ROI report (revenue uplift, cost savings, payback period)
• Operational playbooks (how to maintain and iterate on AI initiatives)
• 12-month AI roadmap (3–5 next initiatives, estimated impact, budget)

Budget: AU$10K–20K (measurement, documentation, planning)

Total 100-Day Budget

• Lightweight approach (quick wins only): AU$50K–80K
• Standard approach (1–2 value initiatives): AU$80K–150K
• Comprehensive approach (2–3 initiatives + compliance): AU$150K–250K

Expected ROI: 2–5x in year 1 (depending on initiatives chosen).

---

Compliance, Risk, and Audit Readiness

AI in healthcare is increasingly regulated. Before you ship, understand your obligations.

Regulatory Landscape (Australia, US, EU)

Australia:

• Privacy Act 1988 (Cth) & Australian Privacy Principles (APPs): Govern patient data handling
• National Health (Pathology) Regulations 2023: If you operate pathology services
• State-based health practitioner regulations: Vary by profession (physio, psychology, etc.)
• AGSM Cybersecurity Essential Eight: Increasingly required by larger customers

US (if you have US patients or operations):

• HIPAA: Governs patient data handling
• FDA guidance on clinical decision support: Increasingly relevant for diagnostic AI
• State-level regulations: Vary widely

EU (if you have EU patients or operations):

• GDPR: Governs patient data handling
• AI Act: Increasingly relevant for high-risk AI (e.g., diagnostic support)

Compliance Checklist for AI Initiatives

Before you launch any AI initiative, confirm:

Data Governance:

• Patient consent obtained (or legal basis established) for AI use
• Data minimisation: Only use data necessary for the AI task
• Data retention: Clear policy on how long data is kept
• Data access: Restricted to authorised staff, audit logging enabled
• Data security: Encrypted in transit and at rest

AI Governance:

• Model documentation: What data was used, how was it trained, what are the limitations?
• Model validation: Independent testing on hold-out data, bias testing
• Explainability: Can clinicians understand why the AI made a recommendation?
• Human oversight: Is there a process for clinicians to override AI recommendations?
• Monitoring: Are you tracking model performance over time? Detecting drift?

Clinical Governance (if clinical AI):

• Clinical validation: Has the AI been tested on real patients? What are the accuracy metrics?
• Risk assessment: What’s the impact if the AI is wrong? (High-risk = diagnostic support; low-risk = scheduling)
• Clinical governance: Is there a clinician responsible for overseeing AI use?
• Incident reporting: Process for reporting AI errors or adverse events

Organisational Governance:

• AI policy: Clear policy on when and how AI can be used
• Training: Staff trained on AI use, limitations, and risks
• Audit trail: All AI recommendations logged and auditable
• Incident response: Process for responding to AI failures or data breaches

SOC 2 & ISO 27001: Worth It?

If you’re planning to sell to larger health platforms, insurers, or enterprise customers, SOC 2 Type II or ISO 27001 certification is increasingly non-negotiable. It’s also a value-creation lever: certified companies command 10–20% higher valuations.

SOC 2 Type II:

• Scope: Security, availability, processing integrity, confidentiality, privacy
• Timeline: 6–12 months (requires 6 months of audit data)
• Cost: AU$30K–80K (depending on complexity)
• Ongoing: AU$10K–20K/year
• Effort: Moderate (requires documentation, process changes, controls)

ISO 27001:

• Scope: Information security management system
• Timeline: 6–12 months
• Cost: AU$40K–100K
• Ongoing: AU$15K–25K/year
• Effort: High (requires comprehensive documentation, risk assessment, internal audits)

Recommendation: If you’re planning a 3–5 year hold and planning to sell to a larger acquirer, start SOC 2 or ISO 27001 in year 1. If you’re planning a quick flip or bootstrapped growth, focus on basic compliance (privacy policy, data security, incident response) first.

For a guided approach, consider engaging a partner like PADISO to conduct an AI Quickstart Audit (AU$10K, 2 weeks) to establish your baseline and create a compliance roadmap.

---

Portfolio Company Benchmarks and KPIs

How do you know if your AI initiatives are working? Here are the benchmarks we see across 50+ allied health portfolio companies.

Operational Metrics

Metric	Baseline	Target (12 months)	Benchmark
No-show rate	25%	15%	10–15%
Clinician admin time	35%	20%	15–20%
Billing cycle time	45 days	25 days	20–30 days
Claim denial rate	8%	4%	3–5%
Patient completion rate	65%	80%	75–85%
Staff turnover	25%	18%	15–20%

Financial Metrics

Metric	Baseline	Target (12 months)	Benchmark
Revenue per clinician	AU$250K	AU$280K	AU$270K–300K
Admin cost as % of revenue	18%	12%	10–15%
Days sales outstanding (DSO)	45	30	25–35
Gross margin	55%	62%	60–65%
EBITDA margin	15%	22%	20–25%

AI-Specific Metrics

Metric	Target	Notes
AI initiative ROI	2–5x in year 1	Depends on initiatives chosen
Time-to-value	<120 days	First initiative should show ROI within 100 days
User adoption	>80%	Clinicians/staff using AI tools in daily workflow
AI accuracy (clinical)	>90%	For diagnostic or clinical decision support
Data quality score	>85%	% of patient records with complete, valid data

How to Measure

Quick metrics (measure weekly):

• No-show rate (by clinician, by time of day)
• Clinician time on admin tasks (self-reported or time-tracking tool)
• Appointment no-shows vs. scheduled

Medium-term metrics (measure monthly):

• Billing cycle time, claim denial rate, DSO
• Patient completion rate, patient satisfaction
• Clinician satisfaction with AI tools

Long-term metrics (measure quarterly):

• Revenue per clinician, gross margin, EBITDA margin
• Staff turnover, patient retention
• Overall practice growth

Pro tip: Set up a simple dashboard (Google Sheets or Tableau) on day 1. Track metrics weekly. This is your north star for decision-making and exit positioning.

---

Exit Positioning: Making AI Visible to Buyers

You’ve built AI capabilities. Now you need to make them visible and valuable to potential buyers. Here’s how.

The AI Story for Buyers

When you’re pitching to potential acquirers (larger health platforms, PE firms, strategic buyers), AI isn’t a technical feature—it’s a value-creation engine. Frame it that way.

The narrative:

1. Problem: “When we acquired [Company], they had a 25% no-show rate, 35% clinician admin time, and 8% claim denial rate.”
2. Solution: “We implemented AI-powered no-show prediction, clinical note automation, and revenue cycle optimisation.”
3. Results: “In 12 months, we reduced no-shows to 15% (+5% revenue), cut admin time to 20% (+15% productivity), and reduced claim denials to 4% (+3% revenue). Total value creation: AU$200K–500K/year.”
4. Repeatable: “This playbook is repeatable across our portfolio. We’ve validated it with 5+ acquisitions.”
5. Defensible: “The AI models are custom-built and integrated into our operations. They’re hard to replicate.”

Data Room Preparation

When you’re in due diligence conversations, have these documents ready:

AI Capability Documents:

• AI roadmap (what was built, when, what’s planned)
• Impact measurement report (revenue uplift, cost savings, ROI by initiative)
• Model documentation (what data, how trained, accuracy metrics, limitations)
• Data governance policy (how patient data is handled)
• Compliance documentation (privacy policy, security controls, audit logs)

Operational Documents:

• Clinician satisfaction survey (AI tools ease of use, impact on workflow)
• User adoption metrics (% of staff using AI tools, frequency of use)
• Incident log (any AI errors, how they were handled)
• Integration documentation (how AI is integrated with EHR, billing, scheduling)

Financial Documents:

• ROI analysis by initiative (revenue uplift, cost savings, payback period)
• Sensitivity analysis (what if adoption is 20% lower? 20% higher?)
• 3-year financial projections (with and without AI)

Valuation Uplift

How much does AI add to exit valuation? It depends on the buyer and the market, but here’s what we see:

Conservative estimate: 10–20% EBITDA multiple uplift (e.g., 6x EBITDA → 7x EBITDA)

• Example: AU$10M EBITDA × 1.15x = AU$1.15M additional value

Moderate estimate: 20–30% EBITDA multiple uplift

• Example: AU$10M EBITDA × 1.25x = AU$2.5M additional value

Aggressive estimate: 30–50% EBITDA multiple uplift (for defensible, high-impact AI)

• Example: AU$10M EBITDA × 1.4x = AU$4M additional value

The uplift depends on:

• Impact size: How much revenue/cost did AI create? (Bigger = higher uplift)
• Defensibility: Is the AI proprietary or easily replicated? (Proprietary = higher uplift)
• Repeatability: Can the acquirer apply the AI playbook to their other assets? (Repeatable = higher uplift)
• Buyer type: Strategic buyers (larger health platforms) value AI more than financial buyers

Due Diligence Talking Points

When the buyer’s technical team asks questions, here’s how to frame your answers:

Q: “How much of your revenue uplift is from AI vs. other factors?” A: “We ran A/B tests. No-show prediction alone accounts for AU$XXX/year. We can show you the control group data.”

Q: “What happens if the AI model degrades?” A: “We monitor model performance weekly. If accuracy drops below 80%, we retrain or roll back. We have a 2-week SLA for model fixes.”

Q: “Is this AI defensible?” A: “Yes. The model is trained on 3+ years of proprietary data. The integration with our clinical workflow is custom. It would take a competitor 6–12 months to replicate.”

Q: “What’s the ongoing cost to maintain this?” A: “AU$XXX/month for SaaS tools, AU$XXX/month for engineering time. We expect to reduce this as we scale.”

Q: “Are you HIPAA/GDPR/Privacy Act compliant?” A: “Yes. We completed a SOC 2 Type II audit in [month]. We have a data governance policy and incident response process in place.”

---

Getting Started: Next Steps

You’ve read the playbook. Now it’s time to move. Here’s how to get started.

Step 1: Assess Your Portfolio (Week 1)

For each portfolio company, answer:

1. What’s the biggest operational pain point? (No-shows, admin overhead, billing delays, etc.)
2. How much data do they have? (Can you access it?)
3. Do they have a CTO or technical leader?
4. What’s their appetite for change?

Score each company on AI readiness (see the checklist above). Prioritise the 1–2 highest-scoring companies for your first 100-day sprint.

Step 2: Engage a Technical Partner (Week 1–2)

You don’t need to hire a full engineering team. Engage a fractional CTO or AI agency that understands healthcare, compliance, and allied health operations.

Look for:

• Experience with healthcare/health-tech (HIPAA, privacy, clinical workflows)
• Experience with allied health specifically (physio, psychology, dental, pathology, etc.)
• Ability to move fast (100-day sprints, not 12-month projects)
• Willingness to take on compliance and audit readiness
• References from other PE firms or health platforms

PADISO, for example, works with PE firms on allied health and health-tech portfolio companies. We offer fractional CTO services (architecture, hiring, vendor calls) and custom AI delivery. If you’re Australia-based, consider Fractional CTO & CTO Advisory in Sydney or AI Advisory Services Sydney. If you’re US-based, consider Fractional CTO & CTO Advisory in Boston. We also offer Platform Development in Philadelphia for healthcare-specific infrastructure.

Step 3: Run the 100-Day Sprint (Week 3–18)

Follow the roadmap above. Mobilise in weeks 1–2, ship first value lever in weeks 3–7, scale and launch second initiative in weeks 8–11, measure and plan in weeks 12–14.

Step 4: Document and Scale (Month 5+)

Once you’ve validated the playbook with 1–2 portfolio companies, document it and scale to the rest of your portfolio. Create a repeatable process:

• AI readiness assessment (1 week)
• 100-day sprint (14 weeks)
• Measurement and planning (2 weeks)

Build a portfolio-level AI roadmap. Allocate a fractional CTO or AI lead to oversee all portfolio companies. This person is responsible for:

• Ensuring consistency in AI strategy and compliance
• Sharing learnings across portfolio companies
• Managing vendor relationships
• Preparing for exit (data room, diligence, valuation story)

Step 5: Prepare for Exit (Month 12+)

As you approach exit (or a secondary transaction), start preparing your AI story. Gather impact data, create a data room, and brief your investment banker on the AI value creation. This is a 2–3 month process.

---

Summary: The Operating Partner Playbook

CapEx vs OpEx decisions in allied health AI come down to one question: Is this a competitive advantage?

If yes, invest (CapEx or hybrid). If no, use SaaS (OpEx).

For most allied health portfolio companies, the hybrid approach wins:

• Use best-of-breed SaaS for commodity tasks (scheduling, billing, document processing)
• Build or co-build custom AI for defensible, high-value workflows (clinical decision support, patient outcome tracking)
• Partner with a fractional CTO or AI agency for architecture, delivery, and compliance

The 100-day sprint is your proving ground. In 14 weeks, you can:

• Identify 2–3 high-value AI initiatives
• Ship the first one and prove ROI (2–5x)
• Build internal credibility and momentum
• Plan the next 12 months

The value levers are clear: no-show reduction (6–12 month payback), clinical productivity (12–18 month payback), revenue cycle optimisation (6–12 month payback), patient outcomes (18–24 month payback).

Compliance is non-negotiable. Start with basic privacy and security controls. Build toward SOC 2 or ISO 27001 if you’re planning a strategic exit.

When you exit, frame AI as a value-creation engine, not a technical feature. Show the revenue uplift, cost savings, and repeatable playbook. You’ll unlock a 10–30% EBITDA multiple uplift—AU$1M–5M in additional exit value.

The playbook is proven. The timelines are realistic. The ROI is tangible. Now go ship.

---

Final Checklist: Before You Start

• Identified 1–2 portfolio companies with strong AI readiness scores
• Secured executive sponsorship (CEO or COO commitment)
• Allocated budget (AU$50K–250K for 100-day sprint)
• Engaged technical partner (fractional CTO or AI agency)
• Scheduled kick-off meeting with portfolio company leadership
• Planned data audit and export (week 1–2)
• Identified first value lever (no-show reduction, clinical productivity, etc.)
• Set up metrics dashboard (Google Sheets or Tableau)
• Created 100-day roadmap (mobilisation, first initiative, scale, measurement)
• Planned compliance baseline assessment (privacy, security, SOC 2 readiness)

You’re ready. Let’s go.