When a journey holds an assessment (knowledge checks, simulations and simulation banks, weighted by its scorecard), its Results stage shows the result of the whole journey first. You can then open each item for its own figures, down to criteria, questions and single attempts.
Runs
A run is one learner's way through the published journey, from its learner link or an LMS package. Restarting the journey starts a new run, and so does opening the journey after a new version is published. Reloading the page in the same browser keeps the same run. The draft preview, review links and the editor's canvas record no runs.
devlin.ai scores every run on its own servers. It uses the run's own simulation attempts, with their evaluations, and its own knowledge check attempts, graded on the server. The journey's scorecard turns them into a score in the same way the learner's journey does. Nothing the learner's browser reports is taken as a score.
- A run is scored with the scorecard the journey had when the run started. Editing the scorecard later changes new runs only.
- Inside a run, each item counts the learner's best attempt at it.
- A simulation that is still being evaluated counts as not complete yet, never as full credit. The run's score updates when the evaluation is saved.
- When an owner or admin overrides a simulation attempt's result, the run is scored again with the adjusted result, and the item's average and attempt list show it. The criteria levels keep the AI's original ratings. A score already sent to an LMS is not changed. See Review or override an evaluation.
- When an owner or admin deletes a simulation attempt, its run is scored again without it.
- A run's score updates as attempts arrive. While a run is still open, a background check scores it again in case an update was missed. If a finished run's score does not show an override, save the override again.
- Results show runs from the day journeys started recording them on the server. Earlier knowledge check attempts stay in the Knowledge check results section below.
Review results
- Open the journey in the editor.
- Select Results.
Journey results shows:
- Learners, Runs, Completed, Passed and Average score. Each learner counts once, by their best run: a pass beats a fail, a complete run beats one still in progress, then the higher score counts. A learner is counted by the name or ID they gave, else by their browser's identifier. A run with no attempt yet, for example from a learner who opened the journey and left, counts in Runs only.
- The scorecard in words, for example "Score: Refund call 75%, Policy check 25%. Pass at 70%."
- Items: each scorecard item with its weight, its average score, how many learners completed it and how many attempts were made at it. The average score and completions come from the same best runs, so the items add up to the journey. The attempts count includes every attempt at the item in every run. An item that is no longer in the scorecard shows "Not scored now".
- Learners: one row per learner with the result and score of their best run, how many runs they made and their score at each item.
Select an item's name to open it. The figures inside also come from the attempt that counts in each learner's best run, and only attempts made inside this journey are included. A simulation used elsewhere keeps its other attempts on its own Results.
- A simulation shows its average score and, for each criterion, the average share of the points earned and how often each level was given.
- A simulation bank shows the same for each simulation the bank drew.
- A knowledge check shows its figures and each question's share of correct answers and option counts. See Knowledge check results.
- Every item lists its newest 200 attempts. For a simulation or a simulation bank, the attempt that counts in a learner's best run is marked "Counts in best run". A knowledge check's question figures come from the counted attempts, but its attempt list does not mark them. A simulation attempt links to the simulation's Results, where the transcript and the full evaluation are.
Before any run, Journey results shows "No runs yet. Runs appear here when learners take the published journey, from its learner link or an LMS package. Previews and review links are not counted."
Results reads the newest 2,000 runs of the journey. When there are more, it says so, and the figures cover those newest runs.
Who sees what follows the knowledge check results rules: any member can open Results, and learner names, IDs and details are shown only to members allowed to see learner identity.
Export to CSV
Export learners CSV downloads the Learners table: one row per learner, with the number of runs, when the best run started, whether it is complete, its score and result, and the score at each item. Identity columns appear only for members allowed to see learner identity.
Learner data
A run stores no learner details of its own. Its learner is whoever its attempts name, so the workspace retention window, merging learners and Delete all data apply through those attempts. Deleting a learner's data also deletes the runs their attempts belong to. The workspace export includes every run with its score. See Data privacy, retention and learner data requests.
Ask devlin and AI tools
devlin, on the Core, Team, Enterprise plans, and AI tools connected over the MCP server in a workspace whose plan includes it and where an owner or admin has turned it on, can read the same results with get_journey_results: the journey's figures, each item's figures, optionally the learners, or one item opened with its criteria or questions and its attempts. Learner identity is hidden from members who may not see it, as in Results.