- Newest
- Most votes
- Most comments
Yes, Amazon Q in Connect now supports 64 languages for agent assistance capabilities, which means you can integrate it with your SharePoint site containing multilingual data files. Customer service agents can chat with Amazon Q in their native language, and Q will provide answers, knowledge article links, and recommended step-by-step guides in that same language.
To set up this multilingual capability, you would need to:
-
Set the appropriate language/locale for Amazon Q in Connect through the Amazon Connect admin website. You can navigate to the AI agents page, create or edit a manual search AI agent, and use the locale dropdown menu to choose the desired language.
-
Ensure your knowledge base (which can include your SharePoint content) is properly set up in Amazon Connect, as this is where Amazon Q will pull information from to assist agents.
-
Configure the necessary permissions for agents to access Amazon Q in Connect, including "Amazon Q in Connect - View" permission to search for and view content, and "Custom views - Access" permission to see step-by-step guides.
The supported languages include Chinese, French, Italian, Japanese, Korean, Malay, Portuguese, Spanish, Swedish, Tagalog, and many others. This multilingual support allows your agents to receive assistance in their preferred language, making it easier for them to help customers efficiently regardless of the language they're working in.
Sources
Set the language for Amazon Q in Connect - Amazon Connect
Amazon Connect launches support for 64 languages for Amazon Q in Connect agent assistance - AWS
Getting started with AWS connect AmazonQ | AWS re:Post
Release notes for Amazon Connect - Amazon Connect
answered a year ago
Short answer: yes, this works for agents typing questions to Q, but there are a couple of things worth knowing before you build it around one mixed-language SharePoint source.
Amazon Q in Connect supports 60+ languages for agent assistance, so an agent can type a question in their own language and get answers, article links, and step-by-step guides back in that language. That's exactly the scenario you're describing, and it's supported. You configure this through the AI agent's locale settings.
The one thing I'd check against your actual workflow: the recommendations Q surfaces automatically from the live conversation transcript are not covered by that full language list. In the "Agent assistance - Proactive Intent Detection based on transcripts" section of the AWS supported-languages documentation, that automatic behavior currently supports only seven languages: English, Spanish, Portuguese, French, Korean, Japanese, and Chinese. So if your agents lean on Q auto-suggesting content as the conversation unfolds, anything outside those seven won't get the same treatment. If they're typing questions themselves, you've got the full list. Worth confirming which of the two your team actually depends on.
One more practical point: the content has to exist in the language the agent is asking in. Q pulls answers from what you've indexed, so you want the German material actually in the knowledge base for German questions, not just English source docs.
On structure, rather than pointing a single knowledge base at one SharePoint site holding every language, I'd split content by language. Amazon Connect added multiple knowledge base support in November 2025, and each Retrieve tool on an AI agent queries a single knowledge base. So you can set up one knowledge base per language and add a Retrieve tool for each. To be clear, that's a recommendation based on how the per-KB retrieval model is designed, not a documented "mixed indexes retrieve worse" rule, so a single mixed-language knowledge base can still work if you'd rather start simple and test it first.
So: multilingual ask-and-answer is supported for agent queries, just make sure your content is present per language, and verify the transcript-based auto-recommendation coverage separately since that list is shorter.
answered 2 months ago
Relevant content
asked a year ago
asked 7 months ago
asked a year ago
- AWS OFFICIALUpdated 2 years ago
