Memory · Data store

Session Summary Store

Data storeMemoryCognition & Memoryarc:SessionSummaryStore

A mid-term memory store of extractive semantic summaries of earlier conversation segments, trading verbatim accuracy for capacity.

Responsibility. Holds compressed summaries of earlier discussion for retrieval when users reference past points.

Also known as: Mid-term memory, Conversation summaries

Direct neighbourhood (hover for relationship types)

Relationships

is read by dependency

Memory Retriever Ch10.1

is written by dependency

Hierarchical History Compressor Ch10.1

Design guidance

SHOULD store key facts (e.g., issue reported, action taken, resolution) rather than every message of a long support thread.

Quantitative guidance

As stated by the sources; verify before use.

Erica keeps semantic summaries for the current session extending 30-40 turns back (Ch10.1).

Classification

Patterns: Hierarchical memory architecture
Quality attributes: Performance efficiency (ISO/IEC 25010)
Risks mitigated: Token budget exhaustion in long conversations

Sources

Ch10.1: T. Nguyen, "Conversational UI," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.1. ISBN: 9798244538229.