An AI roleplay tool with LMS integration and SCORM support should do three things: launch from your LMS with the same login, send a score and a completion status back to the learner's record, and keep the full conversation transcript somewhere you can review it. SCORM can carry the score and pass/fail, but not the conversation itself; xAPI carries richer detail to a learning record store; LTI 1.3 launches the tool and passes the grade back. The most complete option is roleplay that runs natively inside the LMS, which is how Gelato AI works in edzlms: score, rubric feedback, transcript and completion all land in the same gradebook as your quizzes.
Key takeaways
- "LMS integration" and "SCORM support" are not the same claim. Ask which standard, and what data actually comes back.
- SCORM is good for a score and pass/fail. It was not designed to carry a spoken or typed conversation.
- xAPI can record each attempt and each criterion in detail, but you need a learning record store to keep and report on it.
- LTI 1.3 launches the vendor's tool from your course and can post the grade back to your gradebook.
- Native roleplay inside the LMS keeps enrolment, scoring, transcripts and completion in one record.
- Whatever the route, the conversation transcript is personal data. Know where it is stored before you roll out.
- Start with one scenario wired into one existing course, and make it a completion requirement.
What "LMS integration and SCORM support" should mean
Most AI roleplay vendors now say they integrate with your LMS. The phrase covers everything from a link on a course page to a fully tracked activity. For L&D, integration only counts if the practice shows up where every other learning record lives, so a manager can see course progress, quiz scores and roleplay performance for one learner in one place.
There are four ways to get there, and they bring back very different amounts of data:
| Route | How it works | What reaches your LMS | What you lose |
|---|---|---|---|
| SCORM package | The vendor gives you a .zip you upload as a SCORM activity. It usually opens the vendor's roleplay and reports back at the end. | A score and a completion or pass/fail status; sometimes a few interaction records. | The conversation itself and the per-criterion feedback stay with the vendor. |
| xAPI (or cmi5) | The roleplay sends statements ("learner attempted scenario X, scored 7 on objection handling") to a learning record store. | Detailed attempt and criterion data, if your LMS has or connects to an LRS. | Nothing by design, but reporting needs an LRS and someone to build the reports. |
| LTI 1.3 | The roleplay is added as an external tool. Learners launch it from the course with single sign-on. | A grade posted back to the gradebook; enrolment and roles carried at launch. | Transcripts and rich feedback usually stay in the vendor's app. |
| Native activity | Roleplay is an activity type inside the LMS, like a quiz. | Score, rubric feedback, transcript and completion, in the same gradebook and logs. | Nothing, but it only works on the LMS it is built for. |
None of these is wrong. The point is to know which one a vendor means before you sign, because "SCORM support" on a feature list often means the first row.
Does an AI roleplay tool really need SCORM?
SCORM was built for self-contained courses that run in the browser and report a small set of values to the LMS: whether the learner completed or passed, a score, time spent, and a bookmark so they can resume. An AI roleplay is different. The intelligence runs on a server, not in the package, and the valuable output is the conversation and the feedback on it.
So when a roleplay tool "supports SCORM", the package is usually a thin launcher: it opens the vendor's roleplay, waits, and then writes a score and status back. That is useful when your LMS only understands SCORM and you need the practice to count toward completion. It is not a way to get transcripts or coaching data into your LMS.
Ask for SCORM when completion tracking is the requirement. Ask for xAPI, LTI 1.3 or a native activity when managers need to review what was actually said. If you already own SCORM courses, the two work well together: the SCORM module teaches the product or policy, and the roleplay checks that the learner can use it in a conversation. Our guide to AICC vs SCORM vs xAPI vs cmi5 covers the standards themselves.
8 questions to ask any AI roleplay vendor about LMS integration
- Which standard do you use with our LMS: SCORM, xAPI, LTI 1.3, or a native plugin?
- What exactly comes back? Score only, pass/fail, per-criterion scores, the transcript?
- Can the roleplay be a completion requirement in a course, so practice counts toward certification?
- Do learners log in once? Or is there a second account to create and manage?
- Where are transcripts and recordings stored, and for how long? In which country? Can we delete them on request?
- Who writes the scoring rubric? Can we edit the criteria ourselves, and is every attempt scored against the same version?
- What happens to our scenarios and data if we leave? Can we export them in an open format?
- Is there a limit on attempts? The value of roleplay is repetition; pricing that punishes practice defeats it.
- 1Pick one high-value scenario
Start where practice matters most — a common sales objection, a difficult customer call, or a compliance conversation.
- 2Add roleplay as a course activity
In edzlms, add a Gelato AI roleplay to the relevant course just like adding a quiz or assignment.
- 3Set the persona, goal and rubric
Define who the AI plays, what a good outcome looks like, and the criteria it scores against.
- 4Wire it into completion
Make the roleplay a completion requirement so practice counts toward the course and shows on the transcript.
- 5Review the cohort
Watch attempts, scores and transcripts land in the gradebook, and coach the outliers.
How edzlms does it
Gelato AI is edzlms's roleplay engine, and it runs inside the edzlms platform, which is built on Moodle, rather than beside it. A learner opens a course, launches a scored roleplay, and speaks or types through a scenario with an AI persona that responds in real time. When they finish, the score, per-criterion feedback and transcript are written to the same gradebook that holds their quiz and assignment results, and the roleplay can be set as a completion requirement.
Because it is native, there is no SCORM wrapper to maintain and no second system holding the conversations: scenarios, rubrics, transcripts and scores stay in your edzlms or Moodle environment. Your existing SCORM courses keep running alongside it, and edzlms's AI study layer can turn those packages into summaries and practice questions.
For a wider comparison of roleplay tools, see the best AI roleplay tools that integrate with your LMS.
Roleplay bolted on the side
- Separate logins and user lists to maintain
- Scores stay trapped in another tool
- Completion doesn't reach the transcript
- Managers stitch reports together by hand
Roleplay native in edzlms
- One login, one enrolment, one gradebook
- Scores and transcripts sync automatically
- Roleplay counts toward course completion
- Single dashboard for content, quizzes and practice
Pro tip
Make the roleplay a completion requirement, not an optional extra. The moment practice is required to finish the course, attempt rates jump — and so does the score improvement you can actually report.
Frequently asked questions
What is an AI roleplay tool with LMS integration and SCORM support?
It is a conversation-practice tool that launches from your LMS and reports results back to it. SCORM support usually means it can send a score and a completion status. Fuller integrations, through xAPI, LTI 1.3 or a native LMS activity, also bring back per-criterion feedback and, in the native case, the transcript.
Which AI roleplay tools integrate with existing LMS tools?
Many roleplay tools offer a SCORM package or an LTI 1.3 connection that works with most LMSs, including Moodle. Fewer run natively inside the LMS. Gelato AI runs natively in edzlms, which is built on Moodle, so scores, feedback, transcripts and completion all stay in one gradebook.
Can a SCORM package hold an AI roleplay conversation?
Not in a useful way. SCORM reports a small set of values such as completion, pass or fail, score and time. The AI runs on a server, so a SCORM roleplay is usually a launcher that reports a score at the end while the conversation stays with the vendor.
SCORM, xAPI or LTI: which should we use for AI roleplay?
Use SCORM if you only need completion and a score in an LMS that understands nothing else. Use xAPI if you have a learning record store and want detailed attempt data. Use LTI 1.3 for single sign-on and grade passback. Choose a native activity if you want everything, including transcripts, in the LMS.
Can AI roleplay scores appear in the Moodle gradebook?
Yes, through LTI grade passback, a SCORM score, or a native activity. With Gelato AI in edzlms the score, rubric feedback and transcript are written to the gradebook directly, and the roleplay can be a completion requirement.
Does roleplay count toward course completion?
It can if the integration reports a completion status. In edzlms you set the Gelato AI roleplay as a completion requirement, so practice shows on the learner's record and can support compliance or certification.
How is an AI roleplay scored?
The AI evaluates the conversation against a rubric you set, for example clarity, objection handling, empathy and outcome, and returns a score with per-criterion feedback plus the transcript for review.
Where are roleplay transcripts stored?
It depends on the route. With SCORM or LTI tools they usually stay on the vendor's servers. With Gelato AI in edzlms they stay in your own edzlms or Moodle environment. Either way, treat transcripts as personal data and agree retention and deletion rules before rollout.
Sources
- ADL: SCORM and xAPI project pages.
- 1EdTech: Learning Tools Interoperability (LTI).
- Moodle Docs: External tool (LTI) and SCORM activity.
Updated 11 October 2026.
See AI roleplay inside your LMS
We'll run a scored roleplay live inside edzlms, wired to the gradebook and completion, using one of your own scenarios.