# monigatti-rag-to-agent-memory-evolution-2025-11-03

## Veille

RAG to Agent Memory Evolution - Read-write operations - Inference data management - Vector databases - Persistent memory for AI agents - Leonie Monigatti

## Titre Article

The Evolution from RAG to Agentic RAG to Agent Memory

## Date

2025-11-03

## URL

https://www.leoniemonigatti.com/blog/from-rag-to-agent-memory.html

## Keywords

Retrieval-Augmented Generation, RAG, Agentic RAG, Agent Memory, Vector Databases, Semantic Search, Tool Calls, Read-Write Operations, Memory Management, Procedural Memory, Episodic Memory, Semantic Memory, Context Window, Memory Corruption, Data Management, Persistent Learning, Personalization, Conversation History, LLM Agents

## Authors

Leonie Monigatti

## Ton

**Profile:** Technical-educational | Explanatory third person | Analytical-pedagogical | Intermediate

Monigatti adopts a didactic technical-blog style tracing a chronological technological evolution. The pedagogical tone is structured around a clear conceptual progression: RAG → Agentic RAG → Agent Memory. The use of pseudo-code illustrating technical transitions makes the abstractions concrete. Rhetorical questions structure the reasoning ("How to retrieve?" → "Should I retrieve?" → "How is information managed?"). Admissions of nuance ("introduces new challenges such as memory corruption") establish a balanced perspective. Content typical of an ML practitioner explaining architectural patterns to a technical audience familiar with the fundamental concepts but exploring the state of the art.

## Pense-betes

- **RAG (2020)**: single-shot retrieval from external knowledge sources * Offline storage * One retrieval per query * Question: "How to retrieve?"
- **Agentic RAG**: dynamic retrieval via tool calls * The agent determines whether additional information is needed * Multiple retrieval turns possible * Question: "Should I retrieve?"
- **Agent Memory**: read-write operations via tools * WriteTool alongside SearchTool * Persistent learning during inference * Question: "How is information managed?" **Conceptual evolution in 3 stages** 1. **Storage + retrieval** (vanilla RAG) 2. **Agent-decided retrieval** (Agentic RAG) 3. **Full data management** (Agent Memory) **Practical applications of Agent Memory**
- Personalized user experiences via conversation history storage
- Automatic memory creation from important details (preferences, dates, names)
- Multi-source memory systems (procedural, episodic, semantic) **Memory types**
- **Procedural**: know-how, workflows
- **Episodic**: past interactions, context history
- **Semantic**: facts, domain knowledge **New challenges**
- Memory corruption requiring dedicated management strategies
- Validation of write operations
- Memory versioning and conflict resolution
- Privacy and data retention policies **Paradigm shift**
- From retrieval-centered systems to full data management
- Information flows bidirectionally into/out of context windows
- Persistence and learning capabilities fundamentally changed

## RésuméDe400mots

In this technical article, Leonie Monigatti presents the architectural evolution from vanilla RAG (2020) to Agent Memory, tracing the progression in how AI systems access and manage external knowledge, with a focus on the bidirectional flow of information into and out of LLM context windows.

**Vanilla RAG (2020): the foundation layer**

Retrieval-Augmented Generation introduced single-shot retrieval from external knowledge sources. Simple architecture: offline storage + a single retrieval per query. Central question: "How to retrieve?" Semantic search via vector databases augments the LLM with relevant external information. Limitation: deterministic, single-pass retrieval, with no adaptive query refinement.

**Agentic RAG: dynamic retrieval capability**

The evolution introduces tool calls allowing the agent to determine whether additional information is needed. The pseudo-code illustrates the transition:

```
SearchTool available → The agent evaluates relevance → Multiple retrieval turns possible
```

The question shifts: "How to retrieve?" becomes "Should I retrieve?" The agent autonomously decides when and where to retrieve information. Retrieval becomes more strategic, contextual, and iterative. But operations remain read-only: information flows only into the context window.

**Agent Memory: full data management**

"The next logical step after the evolution from vanilla RAG to Agentic RAG." Introduces a WriteTool alongside the SearchTool. Major paradigm shift: read-write operations. The question becomes: "How is information managed?"

The pseudo-code shows the transformation:
```
SearchTool (read) + WriteTool (write) → Bidirectional information flow → Persistent learning
```

Information flows in both directions: not only retrieval, but also storage and modification during inference. Agents' persistent learning capabilities are fundamentally changed as a result.

**Demonstrated practical applications**

**Personalized user experiences**: storing conversation history ensures continuity across sessions. User preferences and interaction patterns are persisted.

**Automatic memory creation**: the system extracts and stores important details (preferences, dates, names) without an explicit user command. Proactive memory management.

**Multi-source memory systems**: architecture supporting distinct memory types:
- **Procedural memory**: workflows, know-how
- **Episodic memory**: past interactions, context history
- **Semantic memory**: facts, domain knowledge

The separation enables specialized retrieval strategies per memory type.

**New challenges introduced**

The article, balanced in its treatment, highlights the challenges:

**Memory corruption**: write operations can introduce errors and outdated information. Validation strategies are needed.

**Management complexity**: versioning, conflict resolution, and retention policies become necessary. More power = more complex governance.

**Privacy**: persistent storage raises questions of data retention, consent, and the right to be forgotten.

**Paradigm shift summarized**

The evolution represents a fundamental shift from retrieval-centered systems to full data management. RAG retrieved knowledge, Agentic RAG decided when to retrieve, Agent Memory manages the entire knowledge lifecycle.

Key quote: "Agent memory represents paradigm shift from retrieval-focused systems to comprehensive data management."

Framework progression: static augmentation → dynamic retrieval → persistent learning. Each stage builds on the previous capabilities by adding a layer of autonomy. Agent Memory enables agents to learn from interactions, build knowledge bases, and personalize responses based on accumulated experience. The transformation from a retrieval tool into a data management platform fundamentally redefines LLM agent architecture.

## GrapheDeConnaissance

- Leonie Monigatti —publie→ analyse de l'évolution RAG vers Agent Memory (DOCUMENT, 0.98)
- Agentic RAG —est_basé_sur→ RAG (TECHNOLOGIE, 0.97)
- Agent Memory —est_basé_sur→ Agentic RAG (TECHNOLOGIE, 0.97)
- Agent Memory —utilise→ opérations lecture-écriture (CONCEPT, 0.95)
- RAG —utilise→ vector databases (TECHNOLOGIE, 0.93)
- Agentic RAG —utilise→ tool calls dynamiques (CONCEPT, 0.94)
- Agent Memory —permet→ apprentissage persistant (CONCEPT, 0.92)
- Agent Memory —utilise→ mémoire procédurale (CONCEPT, 0.9)
- Agent Memory —utilise→ mémoire épisodique (CONCEPT, 0.9)
- Agent Memory —utilise→ mémoire sémantique (CONCEPT, 0.9)
- Agent Memory —permet→ corruption mémoire (CONCEPT, 0.88)
- gestion complète des données —remplace→ systèmes centrés récupération (CONCEPT, 0.92)

---
Canonical: https://www.thekb.eu/en/fiches/monigatti-rag-to-agent-memory-evolution-2025-11-03/