Customer Memory Agent

Remembers each customer’s history across conversations.

Industry	Level	Difficulty	Recommended mode
Customer support	4 of 5	Expert	flat

Coming soon

A complete, runnable Customer Memory Agent reference implementation is in progress. The architecture and core code below already work with SrvDB 0.3.

The problem

• Customers repeat themselves to every agent.
• History is spread across systems.
• Context is lost between channels.

How SrvDB powers Customer Memory Agent

1. One SrvDB database per tenant, or per customer segment.
2. Conversation summaries are stored and recalled by meaning.
3. Screenshots matched with image embeddings.
4. Old data removed with delete and compact().

What changes

Today	With Customer Memory Agent on SrvDB
Customers repeat themselves	Context recalled automatically
Channel silos	One memory
Retention handled manually	Delete and compact by policy

How it works

1. Summarise conversations.
2. Store per customer.
3. Recall on contact.
4. Update after each case.
5. Apply retention.

SrvDB features it uses

• Several databases: each database is a directory, so different data types or tenants stay separate.
• upsert: inserts new items and replaces changed ones, so the knowledge base stays current.
• delete and compact(): remove items immediately from search, then reclaim disk space.
• Crash-safe persist(): after it returns, data survives a crash or power loss and reopens exactly as committed.

Use cases

• Personalised support
• Escalation context
• Account managers’ briefings
• SaaS tenant isolation
• Voice transcript recall
• Retention compliance

Core code

embed() is your embedding model (for example sentence-transformers, or a function that turns your records into feature vectors).

import json
import srvdb

db = srvdb.Database("./customer-memory-agent-db", dimension=384)

def ingest(items):
    """items: list of (id, text, metadata dict)"""
    db.upsert(
        ids=[i for i, _, _ in items],
        embeddings=[embed(t) for _, t, _ in items],
        metadatas=[json.dumps(m) for _, _, m in items],
    )
    db.persist()                      # crash-safe checkpoint

def recall_customer(context, k=5):
    hits = db.search(embed(context), k)
    return [(json.loads(db.get(i)), round(score, 3)) for i, score in hits]

Limits and honest notes

• Respect privacy and retention rules.

Build roadmap

1. Prototype: load a sample of real data with flat mode and check that searches return what an expert expects.
2. Measure: if you move to an approximate mode, compare its results with flat on the same queries (recall).
3. Pipeline: add continuous upsert and decide when to call persist().
4. Harden: test restarts and crashes, add logging, review every output path with domain experts.
5. Deploy: ship the database directory with the application or service.

All Customer support projects

Level	Project	What it does
1 · Beginner	Knowledge Base Search	Customers and agents find the right article from their own words.
2 · Intermediate	Helpdesk Answer Bot	Drafts replies from similar solved tickets and articles.
3 · Advanced	Ticket Deduplicator	Finds duplicate tickets and emerging issue clusters.
4 · Expert	Customer Memory Agent (this page)	Remembers each customer’s history across conversations.
5 · Frontier	Offline Support Kiosk	A self-service assistant that works without internet.

See also: Use cases · Live examples (coming soon)