Compare AI chart generators against my checklist

Task: Design a one-hour test and a scoring sheet that lets me compare AI chart generators on my own data. Context: I want AI to make charts for [kind of work, for example weekly reports]. My data is [type of data] with about [number] rows, and it contains [public, internal or sensitive information]. The tools I am considering are [tool names, or "not chosen yet"]. My must-haves are [must-have 1] and [must-have 2]. Rules: - Do not state features, prices or accuracy of any named tool. Build the test so that I find out by trying. - Include three test tasks of rising difficulty on a sample of my data with the sensitive parts removed, and say how to remove them. - Give a scoring sheet with six criteria scored from 1 to 5, including whether the numbers in the chart match my data when I check five of them by hand, and how the tool handles a wrong or unclear request. - Add a rule for stopping: what result in the first test task means I drop a tool. Result: The three test tasks in one line each, the scoring sheet as a table with criteria and what each score means, then the stopping rule in one sentence.