Reading your results
How to read the prose engine feedback after you submit an exercise.
After you submit an exercise, the prose engine grades your text and shows your results. Here is how to read them.
The headline metrics
The headline metrics at the top of your results show the two or three metrics most relevant to the skill this exercise targets. These are the numbers that matter most for what you were practicing.
The narrative summary
Below the headline metrics is a short, plain-English description of your results. It is assembled directly from your metric scores. No AI writes it, so the same submission always produces the same summary.
The metric tabs
The metric tabs (Readability, Structure, Vocabulary, Style) hold the full set of scores for your submission. Each tab shows the metrics in that category. What each metric means defines every one of them.
What each metric card shows
Each card displays:
- The metric name
- Your score
- A status badge: On Target, Off Target, or Far Off Target, set against the benchmark for your writing type
- A progress bar showing where your score falls relative to the benchmark
- Trend text, if you have prior submissions in this writing type, showing how your score has changed
Benchmarks
Benchmarks are set per writing type. What counts as on target for fiction is different from what counts as on target for technical writing or poetry. The benchmark shown on each metric is the target for the writing type you selected when you submitted. There is more on this in Writing types and benchmarks.
The info button
The small circle icon next to each metric name opens a detail view: the metric's formula, its research source, and a note on how to read it for your writing type.
"From your text"
At the bottom of each metric tab is a "From your text" section. It shows observations tied to specific metrics in your submission. These are deterministic, drawn straight from your scores, not generated by AI.
On Target does not mean perfect
It means your score is within the expected range for your writing type. Some exercises are built to push you toward the edge of the benchmark on purpose. Being off target is information, not a verdict.
Still stuck? Head back to Support to report a bug or reach the team.