When Polymorphism says it uses AI to coordinate home care, we get two kinds of reactions: families who are excited about faster matching and less paperwork, and families who are concerned that a computer is making decisions about their parents. Both reactions make sense — and both deserve a straight answer.

This guide explains exactly what AI does (and does not do) in the Polymorphism platform, how we protect your privacy, and why AI-assisted coordination leads to better care outcomes than the traditional phone-and-spreadsheet model — without replacing the human judgment that matters most.

This article describes Polymorphism's own platform architecture. AI capabilities in home care vary widely across providers. Not all AI claims from care providers are equivalent — ask specific questions about what the AI is actually doing.

The Problem AI Solves in Home Care

Home care coordination is fundamentally an information problem. A coordinator has to know:

What does this client need, exactly? (ADLs, medical complexity, schedule)
Which caregivers have the right skills, credentials, availability, and proximity?
Who has the right language and cultural fit?
Which caregiver profiles have reliability signals consistent with this case?
Is the caregiver's CRC and certification current?
Is this case CDHCI-billable, and if so under which type?
What happened during each visit, and did anything flag for follow-up?

In a traditional model, a human coordinator manages this across dozens of active cases using a combination of memory, spreadsheets, phone calls, and email. It works until it doesn't — which is usually when case volume scales, a coordinator is sick, or an urgent case arrives at 4pm on a Friday.

AI does not eliminate the coordinator. It gives them structured data and ranked suggestions instead of an undifferentiated pile of information — freeing them to focus on judgment calls, not information retrieval.

The 8 AI Agents in the Polymorphism Platform

Polymorphism's platform is built around a set of specialized AI agents, each responsible for a specific part of the coordination workflow. Here is what each agent does, in plain language:

1. Intake Agent

The Intake Agent guides families through the care assessment questionnaire and structures their responses into a standardized care brief. When you answer questions about your family member's ADL needs, schedule, language preferences, and funding situation, the Intake Agent converts those answers into a structured data format that every other agent can work with.

What it does NOT do: make eligibility decisions, give clinical advice, or assess medical need. It captures what you tell it and structures the information.

2. Match Agent

The Match Agent compares the care brief against the caregiver registry and ranks caregivers by fit. It uses a combination of structured criteria (skills, certifications, availability, zone) and semantic embeddings (which capture nuanced information like "five years of dementia care experience" in a way that goes beyond keyword matching).

What it does NOT do: make final match decisions. It produces a ranked shortlist. A human coordinator reviews the shortlist and approves recommendations before they are shown to the family.

3. Schedule Agent

The Schedule Agent handles the scheduling layer — taking approved matches and available time slots, detecting conflicts, and producing shift assignments that respect caregiver availability windows and drive-time constraints. It also sends shift confirmations and reminders to caregivers.

What it does NOT do: override caregiver availability or assign shifts without confirmation. The caregiver confirms each shift.

4. Care Plan Agent

The Care Plan Agent maintains the client's care brief over time — updating it when intake information changes, flagging inconsistencies, and producing formatted care plan summaries for family members or AHS case managers. It also creates structured handoff notes so caregivers have context before a first shift.

5. Billing Agent

The Billing Agent tracks visit events, compiles billable hours, and prepares billing records aligned with Alberta Blue Cross CDHCI requirements. For private-pay cases, it generates invoices based on confirmed visit records from the EVV system.

What it does NOT do: submit claims autonomously. Billing submissions are reviewed and initiated by your coordinator or a designated family admin.

6. Content Agent

The Content Agent handles communication tasks: generating shift summary emails for families, drafting care update notifications, and producing the structured visit notes that caregivers submit after each shift. It works from structured data — actual visit records — rather than hallucinating content.

7. Competitor Intelligence Agent (internal)

This agent is used by the Polymorphism operations team, not by families. It monitors market conditions and helps our team stay current on service offerings and pricing in the Alberta home care market. It does not access any client data.

8. Compliance Monitoring Agent

The Compliance Agent tracks caregiver credential expiry dates, CRC renewal timelines, insurance coverage status, and CDHCI billing compliance signals. When a credential is approaching expiry, it alerts the coordinator and the caregiver — before the expiry creates a gap in coverage or a billing rejection.

What AI Cannot Replace

This is the most important part of this guide. AI handles information processing and structured coordination tasks well. There are several things it cannot and should not do:

Clinical assessment and care planning — AHS nurses, OTs, and case managers determine what care is needed. AI coordinates the delivery; it does not assess the need.
Judgment about family dynamics— the "right" caregiver for a family is not just a credentials match. Personality fit, communication style, and trust matter. Our coordinator review step exists specifically to apply human judgment here.
Response to complex emergencies in the home — AI cannot call 911. The caregiver is the human on the ground, and their training and judgment are what matter in a crisis.
Advocacy— when families need someone to push back on an AHS decision or negotiate with an insurer, that requires a human coordinator who can represent the client's interests.
Grief and emotional support — care coordination intersects with some of the hardest moments in family life. No AI agent handles that. Your coordinator does.

Privacy and Data Protection

We know that the biggest concern families have about AI in health care is privacy. Here is exactly how we handle your data:

What We Collect

We collect information you provide during intake (care needs, schedule, family contact info), information caregivers provide during onboarding (credentials, availability, skills), and structured visit records (time-stamped check-in/check-out, shift notes). We do not collect passive surveillance data, biometric data, or home camera footage.

How It Is Stored

All client data is stored on infrastructure hosted in Canada, subject to PIPEDA (federal Personal Information Protection and Electronic Documents Act) and applicable Alberta privacy legislation including the Health Information Act (HIA) where applicable. Data is encrypted at rest and in transit.

Who Has Access

Your care brief and visit records are accessible to: you (and family members you authorize), your assigned caregiver (for their active cases only), your Polymorphism coordinator, and AHS case managers you explicitly authorize. We do not sell client data. We do not use client health information to train AI models.

AI Model Access to Your Data

The AI agents process your data to generate recommendations and automate coordination tasks. They do not retain personal information across sessions. Recommendations are generated in-context and then discarded — the system stores the structured output (a match recommendation, a schedule entry), not the full data context used to generate it.

What AI-Assisted Coordination Actually Changes

For families, the practical difference between AI-assisted and traditional coordination:

Speed

Traditional agency matching: 5–10 business days from intake to first shift (phone tree model). Polymorphism: 24–72 hours from completed intake to caregiver introduction. For post-hospital discharge cases, this difference is measured in hospital days and risk.

Consistency

In a traditional model, your care coordination quality depends heavily on which coordinator picks up your file — whether they are experienced, whether they know the caregiver roster well, whether they have capacity that week. AI-assisted coordination delivers consistent matching quality regardless of which human coordinator touches the case.

Transparency

Our match recommendations include structured rationale: why this caregiver ranked highly, what skills match, what geographic factors were considered. You can evaluate the recommendation rather than just accept whoever an agency assigns.

Continuity

When a caregiver leaves or needs replacement, the care brief is already structured and current. A replacement can be matched from the same brief in hours — rather than repeating the intake process from scratch.

Common Questions About AI in Home Care

Is AI diagnosing my family member?

No. AI in Polymorphism's platform does not diagnose, assess medical need, or make clinical decisions. It processes structured information you provide and helps coordinate the delivery of care services. All clinical decisions remain with qualified health professionals.

What if the AI makes a bad match?

Every match recommendation is reviewed by a human coordinator before the family sees it. The coordinator can accept, modify, or reject the AI's suggestions. If a match does not work out after the first shift, we re-match at no additional charge using the updated feedback.

Can I opt out of AI involvement?

You can request a fully human-coordinated workflow on any Polymorphism plan. This typically means slower matching timelines, but the option exists. Speak with your coordinator if you have concerns about specific AI functions.

Does Polymorphism comply with Alberta's health privacy laws?

Yes. We operate under PIPEDA and applicable Alberta privacy legislation. Where client health information is involved, we comply with HIA requirements. Our Privacy Policy (available at polymorphism.agency/privacy-policy) provides full details.

The Bottom Line

AI in home care is not about replacing caregivers or making automated decisions about vulnerable people. It is about removing the coordination friction that currently translates into 3-week waits, missed shifts, and information that falls through the cracks. The caregiver who shows up at your parent's door is a trained, credentialed human professional. AI just made sure the right one showed up faster, with context, and with a backup plan already in place.

If you have specific questions about how our AI systems work, we welcome them. Speak with a coordinator at any time — that conversation is always human-to-human.

AI in Home Care: What Alberta Families Need to Know