Neo supports AI-powered dataflow generation and editing using Aira Coder. This allows you to build workflows simply by describing your use case in natural language. Whether you're creating a new dataflow or updating an existing one, the AI interprets your intent and generates the necessary dataflow automatically. You can perform the following using Aira Coder:

Create and modify dataflow (Draft dataflow)
Describe dataflow - Applicable for both draft and approved dataflows
Ask documentation-related queries/help on Neo

The feature is available by default along with Neo.

How It Works: AI Processing Flow

Prompt Enrichment (RAG):
The system enhances your natural language prompt with additional context and examples using Retrieval-Augmented Generation (RAG).
AI-Generated Draft:
The LLM (Large Language Model) translates your prompt into a draft dataflow with all necessary blocks.
Validation Layer:
A system-level validator checks for:
- Missing or misconfigured blocks
- Empty or incomplete nodes
- Syntax issues in script blocks
Auto-Explanation + Preview:
Once validated, the draft dataflow appears in the UI along with a clear, step-by-step explanation of what each block does.
Interactive Refinement:
You can:
- Accept and render the dataflow
- Request changes via a new prompt
- Discard and regenerate

What You’ll See on the UI

Block-by-block Explanation:
Each block includes a short summary of its role in the dataflow.
Preview of Generated Dataflow:
Instantly view the structure of the proposed dataflow before deploying.
Default Error Handling:
AI automatically includes a fallback error-handling block to ensure graceful execution.
Prompt Revisions:
Easily update or refine your prompt to adjust the logic.