# Opa

> Opa is an AI coworker that embeds inside SaaS products. Customers describe what they want; Opa executes it. It replaces help articles, onboarding tours, and traditional support chatbots with a chat surface that does the work.

Opa learns the host product's APIs, data model, and features automatically — no manual tool-writing — and runs playbooks authored in plain English by product, GTM, and CS teams. The end user, who is the SaaS company's customer, talks to Opa inside the product they already use; Opa interprets the goal, calls the right tools, and reports back with the outcome.

The distinction Opa makes: traditional support chatbots return links to documentation; Opa returns finished work.

## Who Opa is for

Opa is sold to product, engineering, and GTM leaders at B2B SaaS companies who are trying to:

- Reduce CS ticket volume by letting customers complete work themselves
- Ship AI-native features without a multi-quarter engineering project
- Improve activation, retention, and feature discovery without rebuilding onboarding
- Avoid being disintermediated by general-purpose AI assistants that talk to their product on the user's behalf

It is not a consumer product, not a standalone chatbot, and not a developer tool for building agents from scratch.

## How it works

- **Connect your product.** Opa ingests everythign about your product and auto-generates the tool surface. Authentication and permissions are inherited from the host product. 
- **Design playbooks.** Product, GTM, and CS teams write plain-English playbooks describing the outcomes customers want. Playbooks are versioned, observable, and editable without engineering involvement.
- **Embed the coworker.** A React SDK component is dropped into the host product's frontend. The chat surface is themed to match the host product's design. 
- **Observe and iterate.** Every agent session has a shareable session ID and is fully replayable end-to-end through the Session Inspector.

## What Opa replaces

- Help articles users do not read
- Onboarding tours users skip
- Support chatbots that return documentation links instead of executing actions
- Custom AI features that take engineering teams quarters to build internally
- The growing share of customer work that is silently moving to general-purpose AI assistants outside the product


## Key product capabilities

- React SDK for frontend embedding
- Automatic updates
- Plain-English playbook authoring (the no-code surface for non-engineers)
- Session-level observability with full replay
- Theming and authentication inheritance from the host product

## Contact

- Site: https://opa.chat
- For partnerships, demos, and enterprise inquiries: contact through the site