> ## Documentation Index
> Fetch the complete documentation index at: https://aletyx.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Decisions + GenAI

> The Model Context Protocol (MCP) pairs deterministic Drools and DMN decisions and Kogito processes with LLMs like Claude for auditable, business-owned AI outcomes.

The **Model Context Protocol (MCP)** is Aletyx's framework for combining
traditional business automation — rules and processes — with advanced AI models
like Claude. Deterministic business logic governs the decision; the LLM handles
the language and context. You get the best of both worlds: AI flexibility with
auditable, business-owned outcomes.

## Why rules and processes still matter in the AI era

Business rules and process automation remain superior for core enterprise needs:

* **Deterministic execution** — identical inputs always produce identical
  outputs.
* **Microsecond-level speed** at high volume.
* **Full auditability** — every outcome records which rules fired and why.
* **Business ownership** — analysts own the logic, not a black-box model.

MCP lets you keep these guarantees while still leveraging AI where it adds
value.

## How MCP works

MCP is an orchestration framework with bidirectional communication between rules
engines, process automation, and AI services. It spans three layers:

<CardGroup cols={3}>
  <Card title="Decision layer" icon="diagram-project">
    Deterministic logic in Drools and DMN.
  </Card>

  <Card title="Process layer" icon="arrows-spin">
    Orchestration with Kogito.
  </Card>

  <Card title="AI integration layer" icon="robot">
    Standardized connectors and prompt management for LLMs.
  </Card>
</CardGroup>

Its core components include a Rule-Governed AI Gateway, a Context Provider, a
Prompt Management System, a Response Processor, and an Integration Framework.

## Real-world applications

* **Dynamic customer service** processes that adapt to the conversation while
  staying within policy.
* **Compliant document processing** with full auditability.
* **Intelligent underwriting** that pairs AI assessment with governed decision
  rules.

## Next steps

<CardGroup cols={2}>
  <Card title="Connecting DMN to LLMs with MCP" icon="plug" href="/docs/faq/products-features/dmn-mcp">
    How DMN models are exposed to LLMs automatically.
  </Card>

  <Card title="Decisions guide" icon="diagram-project" href="/docs/guides/decisions">
    Author, test, and deploy decision models.
  </Card>

  <Card title="Decision Control" icon="shield-check" href="/docs/decision-control/overview">
    Enterprise governance for AI-integrated decisions.
  </Card>
</CardGroup>