# Conversation Memory Overview

Store and reuse contact context to improve continuity between calls and campaigns. Read the Memory guide in Hello Alex documentation.

Canonical HTML: https://docs.helloalex.ai/article/memory-overview

Path: [Documentation](/) / [Memory](/memory)

## Overview

Memory improves continuity because prior interaction context can influence current conversation quality.

## How it works

- Stores retain selected details per contact, making repeat interactions more personalized and efficient.
- Read and write settings define when context is retrieved and persisted, balancing relevance with data governance.

## Use cases

- Recall campaign continuity
- VIP follow-up personalization

## Tips

- Good to know: memory is most impactful for follow-up and retention workflows rather than first-touch outreach.

## Memory Stores

Stores organize persisted context so retrieval stays aligned to business use case.

Define one store per high-impact workflow and document intended variable usage.

- Each store groups a memory domain, allowing teams to separate contexts like recalls, support, or high-value clients.
- Attaching the right store at launch ensures retrieved context is relevant to the current objective.
- **Store name:** Selected on launch workspaces.
- Good to know: separating stores by workflow reduces cross-context noise during retrieval.
- Segmented memory strategy
- Workflow-specific context retention

## Store Detail

Store detail visibility supports quality checks and policy-aligned retention behavior.

Audit sample records and adjust retention to match your governance requirements.

- Record counts and retention settings indicate whether context accumulation matches expectations.
- Sample record inspection confirms that captured variables are meaningful and safe for future reuse.
- **Retention:** How long records are kept.
- Good to know: retention policy should be reviewed with compliance owners before scaling memory usage.
- Compliance review
- Data quality verification

## Sync Settings

Sync timing determines when memory influences conversations and when new context is retained.

Enable one sync mode at a time and verify behavior through controlled end-to-end tests.

- Read-on-launch pulls prior data into active interactions, while write-on-completion persists newly extracted outcomes.
- Careful sequencing prevents stale or premature context from shaping live responses.
- **Write on completion:** Persists variables after calls.
- Good to know: testing with a small audience first helps confirm expected read/write flow before broad rollout.
- Incremental memory rollout
- Behavior verification before scale

## Next step

Create a dedicated memory store for one follow-up use case and test read/write behavior in a pilot batch.

## Related documentation

- [More Memory guides](/memory)