How Ron360 Reports builds a report
Ron360 Reports works differently from a report designer. There is no canvas, no field picker, and no layout palette. You describe what you want and an assistant builds it. Understanding what happens between those two points explains most of what is otherwise surprising about the feature.
The three stages
Fleet360 describes the pipeline in the builder itself: the AI will plan the queries, fetch the data, and render an HTML report you can save or print.
Plan. The assistant interprets your description and works out what to ask the database, which entities are involved, how they join, and how rows should be grouped and aggregated. This is where follow-up questions come from.
Fetch. It runs the planned queries against your fleet data, within your permissions and facility access. The assistant cannot reach data you could not reach yourself.
Render. It produces an HTML report, tables with grouping and subtotals, charts, and KPI strips as the content warrants. The report stays live: it can be re-run against current data, re-filtered, shared, and scheduled.
Generation and rendering are separate stages and fail separately. "Generating report…" means planning and fetching; "Rendering report…" means drawing the result.
Why you ask for things instead of clicking them
Several capabilities that would be toolbar controls in a traditional report designer are decisions the assistant makes while building. Page orientation and invoice-style layouts are selected by the report definition and rendering engine rather than a toolbar control. The conversation is the control surface: "Make this landscape" or "lay this out like an invoice" are valid instructions.
The absence of a control is not the absence of a capability.
What the assistant knows
The assistant is grounded in your fleet's own data and descriptions of the Fleet360 data model. Fleet360 also publishes curated, governed knowledge into Ron360, including terms, business rules, metric definitions, and approved reference material.
What this means in practice
Be specific about the period, scope, and row. Supplying the date range, facilities, and row meaning up front usually produces a usable first draft.
Expect to iterate. Refining by conversation is the normal path.
A vague prompt produces nothing rather than something wrong. The assistant declines rather than inventing a plausible result.
Results respect your access. Two people can see different rows. Scheduled email delivery is generated with the schedule owner's access, see Schedule report delivery.