Generate drafts from your website
Scan supported public pages and use their content to create a first set of articles. Review policies, prices, dates and product details before approving the drafts.
Turn what your business knows into help customers can find, answers agents can trust and knowledge Message AI can use. Create, review and publish from one place.
Build your knowledge and choose where it is published. AI and connected-source availability follow your workspace plan.Published answers, ready for your customers.
Your best answers may be on the website, in a document or inside an agent’s last helpful reply. Bring them into a maintained knowledge workflow.
Scan supported public pages and use their content to create a first set of articles. Review policies, prices, dates and product details before approving the drafts.
If you do not have a website, the onboarding interview asks about your business and follows up based on your answers. Start with short responses, then refine the generated content.
Supported document, spreadsheet and store connectors can supply knowledge. Choose the sources or ranges, inspect the preview and approve what belongs to your business.
Make the everyday questions easy to find. Keep your public help content and internal guidance organized without making everything public by default.
Use collections and clear titles to group questions by the customer’s goal. Delivery, returns and getting started make better entry points than your internal department codes.
Draft, published and archived states make the article lifecycle explicit. Review changes before publication and keep restricted material out of public content.
Configure your help-center identity, logo and theme. Use the provided hosted address or set up a supported custom domain and complete its DNS verification.
A knowledge base should evolve with the business. Review what customers ask and use real support answers to fill useful gaps.
Completed support conversations can suggest reusable questions and answers. Decide whether an answer represents company policy, edit it and approve it before reuse.
Use available coverage and unanswered-question workflows to identify missing material. Improve the article or source behind the answer instead of patching the same response repeatedly.
Revisit articles when products, hours or policies change. Source and publication information help reviewers understand where an answer came from and whether it still belongs in the knowledge base.
The content lifecycle matters as much as the first draft.
Find content by title and text. Agents can get to the relevant answer without scrolling through a long folder of documents.
Group related content so customers and editors can navigate the business’s knowledge. Keep common questions close to the relevant policy or product.
See whether an article was written by a person or generated with AI. Reviewers can give generated drafts the attention they need before publishing.
Keep unfinished or suggested answers separate from approved content. A generated article is a starting point for review, not an automatic business promise.
Choose what belongs in public help and what is only appropriate internally. Knowledge access must follow the intended audience and permissions.
Published and approved knowledge can ground Message AI across supported conversations. Update the source when the business changes so your team and AI work from consistent information.
Start with the website or interview, then add your most common policies and customer questions. Bring in supported sources as needed.
Check accuracy, remove private details, choose visibility and place each article in a useful collection. Review AI-generated drafts before publication.
Open the customer-facing help center and ask Message AI representative questions. Improve missing or unclear content and repeat the checks as your business changes.
Build your knowledge and choose where it is published. AI and connected-source availability follow your workspace plan.