What CORA learns from your website

When you create a workspace and give your company website, CORA reads the site so she knows the basics before your first question.

The onboarding research

Right after the workspace is created, you watch CORA research your company. A live card shows her plan and each step as she reads what the website crawl collected and searches the web for general information about the company, its products, customers and competitors. The crawl also reads PDF documents the site links to, such as price lists.

The research runs on the trial that comes with an account's first workspace. A later workspace starts without trial capacity, so CORA does not read its website when it is created.

Review the company profile

When the research finishes, CORA shows what she learned and proposes three company documents:

• Identity: who the company is, its story, mission, voice and values.
• Offerings: what the company sells, its products, services and positioning.
• Audience: who the company serves, its customer segments and competitors.

Each document has a title, a summary, the full text and the web addresses it came from. Edit anything that is wrong, then select Save canonical docs. The saved documents become company knowledge that CORA and her agents read for context. Skip for now takes you to the workspace without saving. Owners and admins can save the profile.

If the research cannot finish, you can continue to the workspace and fill in the profile later.

Facts saved to company memory

The crawl also picks out short facts, such as contact email addresses, phone numbers, the business ID, services and social profiles. Facts CORA is highly confident about go straight into company memory, marked as coming from the website, so chat can use them at once. Less certain facts are not stored, and facts CORA already knows are not added twice. Each fact keeps a link to the page it came from, and you can correct it like any other memory.

Refresh the website

When your public website changes, an owner or admin can open Workspace settings -> Knowledge sources and select Re-crawl under Company website. CORA keeps the current knowledge until the new sync completes, and Recent sync activity lists earlier runs. If no website is connected yet, Add website leads to the workspace details.