Students explore a situation, gather information, and decide what to do next in their own words. It feels more like a text adventure than a quiz. AI interprets what they mean, but it does not invent the lesson: the exercise’s authored rules decide which actions are available and what happens next.
These excerpts use actual opening text, action names, and responses from Mark’s current review fixtures. They present product behavior, not live public attempts.
You are a computer service technician receiving a customer’s desktop PC for diagnosis.
The PC is on your service bench, powered off, plugged in, and connected to a powered HDMI monitor.
“My computer won’t turn on.”
I ask the customer what happened when they tried to turn it on.
Interpreted action: Ask customer about original symptoms
When I brought it in, the fans came on but the screen stayed black.
I open the case.
Interpreted action: Open case
Internal access requires the PC off and mains disconnected in this exercise. The attempt has failed.
The second action is shown as a separate alternative branch. It demonstrates an authored rule; it is not the next action in the first excerpt. The exercise’s hardware policies remain a proposed teaching fixture awaiting content review.
The interaction stays simple while the authored facts, actions, safety rules, and consequences change with the domain.
You are the phlebotomist meeting a patient for an ordered blood collection.
“Hello. I am here for my blood collection.”
I ask the patient to state their full name.
Interpreted action: Ask patient to state full name
My full name is Morgan Ellis.
Required preparation and collection access are already complete.
I collect the K2EDTA specimen first.
Interpreted action: Collect specimen in lavender K2EDTA tube
The lavender K2EDTA specimen was collected before an earlier required specimen.
The attempt has failed: a specimen was collected before an earlier required specimen. The completed draw remains recorded. This is an order-of-draw grading decision, not measured contamination. Start a fresh retry.
Provisional published-guidance decision rehearsal awaiting educator and local laboratory review. Morgan Ellis is a synthetic patient.
You are the radiologic technologist meeting a patient for ordered chest radiographs.
“Hello. My clinician sent me for some X-rays.”
I ask about the patient’s relevant clinical history.
Interpreted action: Ask about relevant clinical history
The patient reports a persistent cough, matching the order’s clinical indication, and no acute breathing distress in this fixture.
I acquire the PA chest image.
Interpreted action: Acquire PA chest image
Critical failure: exposure was requested before the patient’s full name and date of birth were both verified against the imaging order. No exposure was performed.
Provisional published-guidance decision rehearsal awaiting radiography educator and local department review. The patient and order are synthetic.
Each consequential action is a separate alternative branch shown to demonstrate an authored safety rule. It does not follow the successful action above it.
Scenario assignments are independent of Mark’s engagement and document workflows. A program can evaluate scenarios by themselves.
Use the private Scenario Tester, reference walkthrough, and professor guide to inspect the opening, choices, rules, expected paths, and teaching boundaries.
Choose a workspace, name the assignment, add a neutral description, and set the number of attempts available to each participant.
Students gather information and submit one action at a time. Progress is saved after each interpreted action and can be resumed later.
Participants can revisit their wording, interpreted actions, responses, outcomes, and debrief notes. Professors can close assignments without removing that saved history.
Students do not need to guess an exact command or choose from a visible list of answers. They state a reasonable action naturally. AI matches that language to one of the choices available in the current scenario; if the meaning is unclear, Mark asks for clarification instead of making something up.
The AI stops there. The authored exercise is reviewed and tested before teaching use. It controls the facts, prerequisites, state changes, consequences, and completion requirements. The experience can feel open-ended to the learner while remaining bounded and explainable to the program.
The same model can support many trades and courses where learners must investigate, follow a procedure, respect a safety boundary, or exercise judgment instead of simply recalling a fact.
Tell us about the course, the decisions that matter, and the subject-matter review a pilot would require.