Training Your AI

Articles

What an article is, the six article types, every field in the editor, how sections and markets decide where it shows up, what happens automatically when you save, and how to write articles the AI can actually use.

Workflows

What a workflow is and when it beats a policy, walked through with the WISMO (where is my order) example: nodes and blocks, rule versus AI conditions, tool approvals, how the AI picks a workflow, and how to test one before publishing.

Policies

What a policy is, how the AI finds and applies one, every tab in the editor, what Draft, Semi-Automated and Fully Automated actually do, and how ticket rules, validity dates and custom prompts narrow a policy to the right conversations.

Why did the AI answer that way?

How to open the agent run behind any AI reply, read the sources panel that shows exactly which policies, articles, workflows and past conversations were used, turn what you find into the right fix, and use thumbs up and down so your team builds a real list of failures instead of anecdotes.

Communication Style

How Tone of Voice, Emojis, Verbosity, Headlines and the em-dash toggle shape every AI reply, what the default "Balanced" values really do (nothing), when to move from a preset to a custom tone, and what this screen does not control.

Response Methods

What Draft, Semi-Automated and Fully Automated actually do, the difference between sending and closing, every place the setting lives, why the strictest setting always wins, and how to move a policy up the ladder without guessing.

Knowledge Sources

Every source the AI draws on and what each one is for: your articles, policies and workflows, your website crawl, your product feeds, product documents and PDFs, your past conversations, live data from your systems, and the model's own general knowledge of the world.