Context: Interview take-home for Assured Allies — an insurtech company whose core product helps policyholders age in place, reducing insurer claims. The prompt asked for a scheduling solution for policyholder reachouts that starts minimal and is built for progressive enhancement.

Context

Assured Allies embeds with long-term care insurance providers to proactively reach out to policyholders — helping them access interventions (exercise programs, medication management, home modifications) that reduce the likelihood of entering a care facility.

The problem: scheduling those reachouts is done manually with no consistent cadence. There's no logic for when to follow up after an intervention, no automation, and no feedback loop. The result is inconsistent coverage, missed windows, and opportunities to intervene that pass before contact is made.

Reachout timing is a clinical problem, not just an operational one. A policyholder who received a medication management device and hasn't heard back in two weeks is more likely to have abandoned the intervention than one who received a follow-up call in two days. Getting timing right is how you get outcomes.

Personas

Two distinct personas, with different relationships to the scheduling problem:

  • The policyholder — direct recipient of the reachout. Values feeling heard, knowing someone has a plan, and not being overwhelmed. From personal experience caring for older family members: the worst outcome isn't too many calls — it's the sense that nobody is following through.
  • The ally (Assured Allies staff) — the person making the calls. A better scheduling system reduces their manual workload, gives them a prioritized call list, and frees them to focus on the conversation rather than figuring out who to call next.

The scheduling solution serves both: better cadence for the policyholder, better operational tooling for the ally, better clinical outcomes for the insurer.

Why the MVP Isn't Enough — But Start There Anyway

The obvious MVP: a generic two-week cadence for all reachouts, regardless of intervention type. It's consistent, easy to implement, and creates baseline coverage where there is none.

But it's worth being honest about the gaps. Two weeks is wrong for too many interventions.

  • A policyholder who left a voicemail shouldn't wait two weeks for a callback — that's a missed signal of engagement
  • A policyholder who received a medication management device may have a simple setup question; two weeks of silence risks non-adherence
  • A policyholder who received discharge planning education is in an active transition — a two-week delay creates real clinical risk

Generic timing risks lost policyholder motivation, missed intervention windows, and — in some cases — adverse events. That said: start with the two-week cadence anyway. It establishes rhythm, lets the team validate the automation stack, and creates a baseline to measure against. Ship the rubric model next.

The Tiered Model

Interventions are classified into three tiers based on clinical urgency and the time-sensitivity of the follow-up. Tier classification was informed by surveying nurses, PTs, and PTAs — not PM intuition.

Tier 1 — Urgent
2 business days
Active engagement signals or high clinical urgency. Missing the window risks disengagement or adverse events.
Voicemail received
Discharge planning education
Tier 2 — Standard
5 business days
Interventions requiring timely follow-up to support adherence and confirm successful adoption.
Medication management device
Home safety evaluation arranged
Assistive device sent
Tier 3 — Routine
10 business days
Educational and lifestyle interventions with lower urgency — important, but not time-critical.
Exercise program subscription
Chronic condition education
Home optimization coaching

Scheduling Rubric

Full classification of interventions, based on clinical feedback from nurses, PTs, and PTAs:

Intervention Tier
VoicemailTier 1
Educate on discharge planning / servicesTier 1
Arrange home safety evaluationTier 2
Arrange home modificationTier 2
Send assistive deviceTier 2
Send medication management equipmentTier 2
Educate on medication managementTier 2
Set up online caregiver education and supportTier 2
Caregiver coachingTier 2
Home optimization education & coachingTier 3
Subscription to Team Vivo exercise programTier 3
Send vitals log and educational materialsTier 3
Educate on chronic condition managementTier 3
Set up Chronic Disease Self Management CourseTier 3
Educate on bathingTier 3
Educate & coach on home optimizationTier 3
Set up policyholder with transportation optionsTier 3
Create activity calendarTier 3
Health coachingTier 3
Connect with Friendship Line / local council on agingTier 3
Send support materialsTier 3

Automation Stack

The scheduling logic only scales if it's automated. Manual scheduling is the current state — it's where the inefficiency lives.

CRM as the source of truth. The CRM tracks the last intervention per policyholder and calculates the next reachout date based on tier. This populates the ally call list daily — no manual planning required.

Automated outbound calling (Twilio or equivalent). The CRM feeds into an automated cloud-based call center. Each day, the system works through the list, making outbound calls. When a policyholder answers, the call is handed to a live ally immediately.

The handoff is critical. Policyholders who pick up and hear silence — or a robotic opener — will assume scam and hang up. All outbound calls must display Assured Allies caller ID and connect to a live person within seconds of the policyholder saying hello.

Reminder notifications. Policyholders receive an alert (SMS, email, or call — their preference) 4 days before the scheduled reachout and again the day before. This reduces missed calls and gives policyholders a sense of structure.

Secondary contact consent. Policyholders can elect a secondary representative to receive reachouts on their behalf if they're unavailable. Reduces missed contact without requiring repeated call attempts.

Roadmap

Phase 1
MVP
Two-week generic cadence + call automation
Implement automated outbound calling with a flat two-week reachout schedule. Validate the automation stack, measure ally efficiency, and gather baseline data. Run for several weeks before cutting over. Watch for workload spikes — two weeks may be more frequent than the current model, which could stress the team.
Phase 2
Tiered
Tiered scheduling rubric goes live
Replace the flat cadence with the three-tier model. Survey allies and — where possible — policyholders on frequency. Begin the build-measure-learn loop: are Tier 1 windows catching more engagement? Are Tier 3 policyholders feeling under-served?
Phase 3
ML/AI
Data-informed tier refinement
As reachout data accumulates — intervention success rates, time-to-adverse-event, aging-in-place duration — ML models learn which timing windows produce the best outcomes. Tiers are adjusted dynamically rather than manually. The rubric becomes a starting point, not a ceiling.

Success Metrics

Operational

Time per interaction decreases. Successful contact rate increases. Ally call list is auto-generated daily with no manual scheduling.

Engagement

Policyholder reachout coverage increases. Voicemails and missed calls decrease as reminder notifications reduce missed windows.

Clinical

Intervention adherence improves — especially for Tier 1 and Tier 2 cases where follow-up timing directly affects outcomes.

Financial

Over time: policyholders age in place longer, reducing insurer claim frequency. Reduction in premiums paid is the long-term North Star metric.