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

# Gen AI End-to-End Walkthrough

> Walk through an end-to-end Gen AI pipeline from raw data to continuous model improvement.

# Gen AI end-to-end walkthrough

This walkthrough shows how teams commonly operationalize Gen AI using Encord.

***

## Step 1: Ingest and structure data

Centralize documents, text, and multimodal assets with metadata for traceability.

**Docs**

* [Files](/platform-documentation/Curate/index-files)
* [Supported Data](/platform-documentation/General/general-supported-data)

***

## Step 2: Curate for grounding

Filter, cluster, and select trusted sources for retrieval and evaluation.

**Docs**

* [Getting Started with Index](/platform-documentation/Curate/index-getting-started)
* [Embedding Plots](/platform-documentation/Curate/embedding-plots)

***

## Step 3: Collect human feedback

Design annotation workflows for correctness, preference, and alignment.

**Docs**

* [Annotate & Review](/platform-documentation/Annotate/annotate-label-editor/annotate-label-editor-annotate)
* [Ontologies](/platform-documentation/Annotate/annotate-ontologies/annotate-ontologies)

***

## Step 4: Evaluate model behavior

Compare outputs across prompts, datasets, and model versions.

**Docs**

* [Model Evaluation](/platform-documentation/Validation/active-how-to/active-model-predictions-eval)
* [Quality Metrics](/platform-documentation/Validation/label-validation-basics#quality-metrics)

***

## Step 5: Iterate continuously

Feed insights back into curation, annotation, and prompting.

***

## What success looks like

* Hallucinations are **detectable**
* Feedback is **consistent and structured**
* Improvements are **measurable over time**
* Teams iterate **with confidence**