Longitudinal Memory
Configure memory features for AI agents to remember subject context over time
User Repeats Attempt to Ask Questions
If a user interrupts to asks questions more than once, an intervention agent may end the session. This memory is to watch if a user previously tried to ask a question.
Memory Type
RepeatedQuestions
Retention
each_session
Escalation
1x threshold
EHR Link
Not Connected
Daily Interaction Continuity
Remember conversation context from previous interactions for natural dialogue flow
Memory Type
ConversationHistory
Retention
4weeks
Escalation
3x threshold
EHR Link
Not Connected
Symptom Progression Monitor
Remembers reported symptoms over time and tracks severity trends
Memory Type
ConversationHistory
Retention
12weeks
Escalation
4x threshold
EHR Link
Not Connected
Subject Name & Context
Remembers subject's name and personal details for personalized conversations
Memory Type
NameRecall
Retention
lifetime
Escalation
3x threshold
EHR Link
Connected
Glucose Threshold Tracker
Tracks repeated glucose threshold breaches and escalates after pattern detected
Memory Type
ThresholdTracking
Retention
12weeks
Escalation
3x threshold
EHR Link
Connected
Medication Adherence Tracker
Tracks medication compliance over time and identifies patterns
Memory Type
ThresholdTracking
Retention
12weeks
Escalation
3x threshold
EHR Link
Not Connected
1. AI agents access configured memory features
2. Subject interactions are stored with context
3. AI references past events in responses
4. Auto-escalates after threshold count reached
Name Recall
AI remembers and uses subject's name naturally in conversations
Threshold Tracking
Tracks repeated breaches, references last occurrence, escalates after N incidents
Patient Record Link
Integrates EHR data for complete subject context
Transform longitudinal memory data into predictive intelligence across three powerful dimensions:
Per Test Subject
Build personalized predictive models for each individual. ML algorithms identify unique patterns in subject behavior, anticipate adverse events before they occur, and recommend personalized intervention timing. Predict medication adherence, glucose trends, or symptom progression specific to each patient's history.
Across Full Study
Aggregate memory data from all subjects to discover population-level insights. Identify common risk factors, optimal intervention timings, and treatment effectiveness patterns. ML models detect cohort trends, predict dropout risk, flag systemic safety signals, and uncover hidden correlations that drive study outcomes.
Compare Studies
Leverage structured longitudinal data to benchmark across multiple trials. ML identifies which study protocols yield the best outcomes, what subject characteristics predict success, and how interventions perform across different populations. Build a knowledge base that continuously improves trial design and execution.
🚀 Future Capability: These ML features will be available as longitudinal memory data accumulates across your studies.