Introduction
AI roleplay lets learners practice real-world conversations with a virtual avatar before they have those conversations for real. A learner speaks (or types) to an avatar that plays a customer, an employee, a patient, a colleague or any other role you define. When the conversation ends, the AI evaluates it against criteria you set and gives the learner a personal report.
You can use AI roleplay for any type of conversation, for example sales calls, customer support, feedback and performance discussions, onboarding or compliance-sensitive conversations.
AI roleplay has its own section in the platform, called AI roleplay. You create, manage and monitor roleplays there, without needing Creator. This article describes what you can do with AI roleplay, how the pieces fit together and where to find the details.
Requirements and limitations
- AI roleplay is available only if your platform has an active Simulation Pack and the AI roleplay feature is enabled in the Artificial intelligence control panel.
- Learners cannot start a simulation if the Simulation Pack is not active or if no simulations remain for them. Learn more in AI roleplay FAQs and troubleshooting.
- The learner experience can be delivered in many languages, but the roleplay configuration panel is available in English only, regardless of your platform language.
- AI roleplay does not consume AI credits. Usage is measured in simulations.
- AI roleplays are not included in SCORM exports.
- The AI evaluates the conversation transcript, not tone of voice or body language. Write your evaluation criteria with this in mind.
Please note: If you already use AI roleplay in Creator lessons, your existing roleplays keep working. Read AI roleplay in Creator lessons after the standalone release to see what changes.
Why use AI roleplay
Knowing what to say and being able to say it are different skills. Courses and videos build the first, and practice builds the second. Live practice, such as peer roleplays or coaching sessions, works but is hard to repeat, schedule and apply consistently across a large team.
AI roleplay helps close that gap:
- Learners practice on their own time. There is no one to schedule and no audience, so learners can try, make mistakes and try again.
- Every learner gets feedback. The report points to specific moments in the conversation, so learners know what to keep doing and what to change.
- Practice follows your standards. You decide what good looks like by writing the roleplay and the evaluation criteria yourself, instead of relying on generic feedback.
- Managers and subject matter experts stay involved. You can assign reviewers who validate the AI-generated evaluation before a result is final.
- You stay in control of usage. On larger or distributed organizations, simulation pools let you decide which groups or branches can use the simulation balance you purchased.
- Results sit with the rest of your learning data. Learners reach roleplays through courses, so completion appears alongside the other training in the course.
What you can do with AI roleplay
| Capability | What it gives you | Learn more |
| Dedicated AI roleplay section | Create, edit, preview, publish, and duplicate roleplays from the platform navigation. Creator is not required. | Creating and managing AI roleplays |
| Prompt or template start | Describe the conversation you want, or start from a featured template. The AI fills in the roleplay for you to review and adjust. | Creating and managing AI roleplays |
| Customizable avatars | Define who the avatar is: name, role, background and personality (for example agreeable, vague or skeptical, or your own combination of traits). Pick how the avatar looks. | Best practices for writing AI roleplays |
| Three interaction modes | Video (animated avatar), Voice (static avatar image, spoken conversation) or Chat (text only). | Creating and managing AI roleplays |
| Custom evaluation | Use the default rubric, or write your own: qualitative (Good or Developing) for coaching, or quantitative (scored 1 to 5, with optional weights and a pass score) when you need a measurable result. Let AI draft the criteria from your description and documents. | Configuring evaluation, rubrics and reviewers in AI roleplay |
| Human review | Assign managers, groups, branches or individual users as reviewers of learner submissions. | Reviewing AI roleplay submissions |
| Multiple languages | Learners practice in the course language, with the avatar voice and the report in the language they use. | Creating and managing AI roleplays |
| Delivery through courses | Publish a roleplay as a training material and assign it to the courses where you want learners to practice. | Assigning AI roleplays to learners |
| Simulation pools | Reserve parts of your simulation balance for specific groups or branches, so each team has its own allocation. | Managing simulation pools |
How AI roleplay works
First, create a roleplay. In the AI roleplay section, you describe the situation, the learner's goal, the avatar and how the learner will be evaluated. You preview it exactly as a learner will see it.
Next, publish the roleplay. It then becomes available to use.
Then deliver it. When you publish the roleplay, you assign it to a course through a training material. Learners find it in the course like any other content.
When a learner practices, the learner starts the simulation, has the conversation and ends it. Each run counts as a simulation against your balance, or against the simulation pool the learner belongs to.
After the conversation, the AI analyzes the full conversation transcript and generates a report based on your criteria.
Optionally, a reviewer checks the result. If you assigned reviewers, learners submit their attempt and reviewers validate the AI-generated evaluation.
Who can use AI roleplay
| Role | What they can do |
| Superadmins | Everything in the AI roleplay section: create and manage roleplays, assign them to courses and manage simulation pools. |
| Power Users | At release, the AI roleplay section is available to Superadmins only. Power Users do not have access to it. |
| Reviewers | Review the submissions of the roleplays they are assigned to. |
| Learners | Practice the roleplays assigned to them in their courses and read their reports. |
Where to go next
| If you want to | Read |
| Build your first roleplay | Creating and managing AI roleplays |
| Write roleplays that feel realistic | Best practices for writing AI roleplays |
| Define how learners are evaluated | Configuring evaluation, rubrics and reviewers in AI roleplay |
| Get roleplays to learners | Assigning AI roleplays to learners |
| Understand what learners see | Practicing with AI roleplay and reading your report |
| Review learner submissions | Reviewing AI roleplay submissions |
| Control who uses the simulation balance | Managing simulation pools |
| Find answers to common issues | AI roleplay FAQs and troubleshooting |