Senior Drift Intelligence

From 100 daily check-ins to 7 prioritized follow-ups.

Senior Drift Intelligence is a simulation MVP for home care agencies. It generates daily check-in data for 100 elderly users over 30 days, creates a personal baseline from the first 14 days, and flags behavioral drift from day 15 onward.

The business is not AI phone calls themselves. The business is personal-baseline drift detection and care triage: turning a large daily check-in list into a smaller, prioritized caregiver follow-up queue.

Business Concept

Home care agencies need to know where human attention is most useful each day. Subtle changes can be difficult to spot when each person has a different normal pattern. This MVP shows how Senior Drift Intelligence compares each senior against their own baseline and produces family, agency manager, and nurse summary reports.

Potential users include home care agencies, assisted living operators, family caregiver platforms, and remote wellness call centers. The value proposition is operational awareness, reduced alert fatigue, better prioritization of caregiver time, and clearer documentation of non-medical check-in needs.

What the Simulation Does

• Generates synthetic daily call records for 100 seniors across 30 days.
• Tracks call duration, response delay, speech energy, speech rate, sentiment, confusion, repetition, medication confirmation, medication confidence, pain keywords, sleep issues, and appetite issues.
• Builds a personal baseline from days 1-14.
• Calculates a Senior Drift Score from days 15-30.
• Classifies each senior each day as Normal, Watch, Check Today, or Urgent Follow-up.
• Applies escalation rules to reduce alert fatigue.
• Exports family, agency manager, and nurse summary reports, CSV files, and SVG charts.
• Shows business KPIs such as priority compression ratio, estimated caregiver minutes saved, and alert burden score.

Chart Previews

Risk Trend By Day

Seniors By Risk Level

Daily Follow-Up Workload

Individual Senior Drift Timeline

Demo Result: Day 30

• 100 seniors monitored.
• 3,000 simulated records.
• Day 30 Urgent Follow-up count: 2.
• Day 30 Check Today count: 5.
• Priority compression ratio: 14.3:1.
• Estimated caregiver minutes saved: 668.
• Alert burden score: 23.5 / 100.
• 48 daily reports generated.
• 4 charts generated.
• Positioning: non-medical care triage for home care operations.

Why This Is Different

• Not just AI phone calls: the value is the drift intelligence layer after check-ins are completed.
• Personal baseline comparison: each senior is compared against their own first 14 days instead of a generic average.
• Alert fatigue reduction: thresholds and escalation rules keep the highest-priority list smaller.
• Caregiver workflow prioritization: the output is a follow-up queue, not a raw event feed.
• Family, agency, and nurse-specific reporting: the same signal is packaged for different operating roles.

Suggested Buyer

• Primary buyer: home care agencies and visiting care centers.
• Secondary buyer: family caregiver platforms.
• Later buyer: local governments and remote wellness call centers.

Portfolio Docs

• Business model
• SCQA problem statement
• Pilot plan
• Customer discovery questions

Risk Levels

• Normal: No meaningful behavioral drift detected today.
• Watch: Some drift detected; monitor the next check-in.
• Check Today: Caregiver check-in recommended today.
• Urgent Follow-up: Caregiver follow-up recommended today because the score is very high, several major reason codes are present, or priority status continued for two days.

These labels are prioritization aids for care operations.

Thresholds And Escalation

• Normal: Senior Drift Score below 35.
• Watch: Senior Drift Score from 35 to below 65.
• Check Today: Senior Drift Score from 65 to below 85.
• Urgent Follow-up: score at or above 90, score at or above 85 with at least 3 major reason codes, or Check Today or higher for 2 consecutive days.

Run

python3 src/main.py

Outputs

• output/synthetic_call_data.csv: Raw synthetic daily call data.
• output/scored_daily_risk.csv: Daily drift score, risk level, reason codes, and recommended actions.
• output/executive_summary_latest.md: Day 30 business-facing executive summary.
• output/reports/family_report_day_15.md through family_report_day_30.md: Family-facing reports.
• output/reports/agency_manager_report_day_15.md through agency_manager_report_day_30.md: Agency operations reports.
• output/reports/nurse_summary_report_day_15.md through nurse_summary_report_day_30.md: Nurse summary reports.
• output/family_report_latest.md, output/agency_manager_report_latest.md, and output/nurse_summary_report_latest.md: Day 30 report copies.
• output/charts/risk_trend_by_day.svg: Average and 90th percentile drift score trend.
• output/charts/risk_level_counts.svg: Daily count by risk level.
• output/charts/individual_senior_drift_timeline.svg: Timeline for the highest-risk senior on day 30.
• output/charts/daily_followup_workload.svg: Daily count of seniors requiring human follow-up.

MVP Boundaries

This is a simulation with synthetic data and transparent rule-based scoring. A production system would require consent workflows, privacy controls, call recording governance, model monitoring, escalation policies, caregiver feedback loops, and careful validation against real-world operational outcomes.