- They’re available 24/7 to customers all over the world and respond with accurate, standardized information tailored to each query.
- They handle high volumes of inquiries, which reduces wait times and live customer interactions, so your personnel can focus on more complex issues.
How Agent Studio works
Agents come from a prebuilt library. Choose the agent type that fits your use case, and then configure it for your business. Your data stays in your GCS bucket. Agent Studio reads JSONL and PDF content from Google Cloud Storage (GCS) buckets that you connect to the agent. Models are matched to your industry. Select your industry domain, and Agent Studio recommends the generative AI model to use. You can select a different model. Agent Studio doesn’t support custom large language model (LLM) training. Agents embed with a script and one HTML element. Add the agent’s script and<lw-template> snippet to your site, and the agent displays on those pages.
Get started
Quickstart
Create and embed your first agent.
Agent Studio for Commerce and Knowledge Management
Learn how to choose the right agent for commerce and knowledge management use cases, connect it to your data in a GCS bucket, and embed it on your site.
Configure and deploy an agent
This demo shows how to create and deploy an agent in Agent Studio.Product Q&A
Generate frequently asked questions and answers about a product from its documentation.
Conversational Q&A
Answer natural-language questions and follow-ups about a product, grounded in its documentation.
Conversational Q&A FAQ
Combine conversational answers with generated frequently asked questions.
Manage an existing agent
Manage agents
Preview, edit, and delete agents.