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

# Overview of Observe stage

> Learn how to observe your deployed AI capabilities in production using Axiom AI SDK to capture telemetry.

In the Observe stage of the AI engineering lifecycle, the focus is on understanding how your deployed generative AI capabilities perform in the real world. After creating and evaluating a capability, observing its production behavior is crucial for identifying unexpected issues, tracking costs, and gathering the data needed for future improvements.

## Instrument your app

Axiom offers the following approaches to capture generative AI telemetry:

| Instrumentation approach                                             | Language support | Characteristics                                              |
| :------------------------------------------------------------------- | :--------------- | :----------------------------------------------------------- |
| [Axiom AI SDK](/ai-engineering/observe/axiom-ai-sdk-instrumentation) | TypeScript       | Quick setup.<br />Minimal code changes.                      |
| [Manual](/ai-engineering/observe/manual-instrumentation)             | Any              | More involved setup.<br />Full control over instrumentation. |

**Instrumentation with Axiom AI SDK** is the right choice for you if you have a TypeScript app and you want the SDK to capture and send traces with the correct semantic conventions.

**Manual instrumentation** is the right choice for you if you want to use your own tooling or if you use a language other than TypeScript. You need to instrument your app manually to emit traces compatible with Axiom’s AI engineering features.

Both approaches emit identical attributes. This means that all the telemetry analysis features work the same way.

## Review production conversations

Review and Issues let domain experts inspect production conversations, capture human feedback, and group recurring failures into a tracked backlog.

* Use [Review conversations](/ai-engineering/iterate/review) to work through flagged, recent, feedback-driven, errored, and previously reviewed conversations.
* Use [Track issues](/ai-engineering/iterate/issues) to consolidate repeated failures, inspect supporting evidence, and manage issue status over time.

## Visualize traces in Console

Visualizing and making sense of this telemetry data is a core part of the Axiom Console experience:

* A dedicated **AI traces waterfall view** visualizes single and multi-step LLM workflows, with clear input/output inspection at each stage.
* A pre-built **GenAI OTel dashboard** automatically appears for any dataset receiving AI telemetry. It features elements for tracking cost per invocation, time-to-first-token, call counts by model, and error rates.

### Access AI traces waterfall view

1. Click the Query tab.

2. Create an APL query about your GenAI dataset. For example:

   ```kusto theme={null}
   ['otel-demo-genai']
   | where ['attributes.gen_ai.operation.name'] == "chat"
   ```

   [Run in Playground](https://play.axiom.co/axiom-play-qf1k/query?initForm=%7B%22apl%22%3A%22%5B'otel-demo-genai'%5D%20%7C%20where%20%5B'attributes.gen_ai.operation.name'%5D%20%3D%3D%20'chat'%22%7D)

3. In the list of trace IDs, click the trace you want to explore.

4. Explore how spans within the trace are related to each other in the waterfall view. To only display AI spans, click **AI spans** in the top left.

<Frame>
  <img src="https://mintcdn.com/axiom/Xk0oAkpltm6Ivwpw/doc-assets/shots/gen-ai-waterfall-view.png?fit=max&auto=format&n=Xk0oAkpltm6Ivwpw&q=85&s=606ebe2c6e300591f6d75d3404622b72" alt="AI traces waterfall view" width="1758" height="592" data-path="doc-assets/shots/gen-ai-waterfall-view.png" />
</Frame>

### Access GenAI dashboard

Axiom automatically creates the GenAI dashboard if the field `attributes.gen_ai.operation.name` is present in your data.

To access the GenAI dashboard:

1. Click the Dashboards tab.
2. Click the dashboard **Generative AI Overview (DATASET\_NAME)** where `DATASET_NAME` is the name of your GenAI dataset.

[Run in Playground](https://play.axiom.co/axiom-play-qf1k/dashboards/genai.otel-demo-genai)

The GenAI dashboard provides you with important insights about your GenAI app such as:

* Vitals about requests, broken down by operation, capability, and step.
* Token usage and cost analysis
* Error analysis
* Comparison of performance and reliability of different AI models

<Frame>
  <img src="https://mintcdn.com/axiom/Xk0oAkpltm6Ivwpw/doc-assets/shots/gen-ai-dashboard.png?fit=max&auto=format&n=Xk0oAkpltm6Ivwpw&q=85&s=7b7678c587d2fcfc3046d5af2753aedf" alt="AI traces waterfall view" width="2552" height="1668" data-path="doc-assets/shots/gen-ai-dashboard.png" />
</Frame>

## What’s next?

After capturing and analyzing production telemetry, use these insights to improve your capability. Learn more in [Iterate](/ai-engineering/iterate).