AI Components
Recommendations
Statement can include suggested next steps in its reports and explanations, based on the data retrieved for your request. These help connect a finding with something you may want to investigate or act on. For example, if an invoice review shows that a customer’s payment is 45 days overdue, the explanation could highlight it and suggest following up. The recommendation depends on the available invoice dates and payment status. Suggestions are guidance for you to review. Sending a reminder or taking another action requires an explicit request or a workflow you’ve configured. Connecting a data source alone does not mean Statement continuously monitors it and sends proactive recommendations.
Insights
Insights explain what stands out in the data, such as a change over time, an unusual result, or a meaningful difference between categories. For example: “Expenses increased by 20% compared with the previous month.” Recommendations suggest what to do next. An insight might identify higher expenses; a recommendation might suggest reviewing the categories responsible for the increase. One describes the finding, while the other proposes a response. An app can include a written summary that updates with its data, allowing these observations to appear without a new question each time. This needs to be included in the app’s setup-for example, “Add a summary highlighting the biggest changes from the previous month.” It isn’t automatically included in every app.
Chat
In-app chat continues the app's associated conversation and can retrieve further project data. Describe the metric or period you mean; clicking a chart does not automatically pass that selection to chat.
Alerts
Alerts are configured conditional workflows. Define the check, schedule, and destination; connecting a source does not itself activate monitoring.