How to turn a written case study into a scene a class can discuss
Give Vizard Agent the written case and ask it to find a shot for each behaviour the case describes rather than for each noun. It sources the beats separately — distraction, exclusion, a supportive conversation — lays every candidate out on one sheet to check they read as the same class, and closes on question cards so the video ends where the discussion starts.
What is the short version?
A case study reads as a sequence of behaviours: somebody keeps checking their phone, somebody else carries the work, a teacher notices. Searching for the objects in that description — a phone, a classroom — returns stock that illustrates nothing.
- Give Vizard Agent the written case as it stands.
- Ask for a shot per behaviour, not per object.
- Ask for question cards at the end.
What do you need before you start?
The case as written and a sense of what the class is meant to argue about. Vizard Agent builds the visuals from the behaviours in the text, so a case written as a situation works better than one written as a moral.
- The written case. In full.
- The discussion questions. Or ask for suggestions.
- The age group. It sets the casting and tone.
- The length. Usually two or three minutes.
- Any music. A bed changes how the case lands.
What do you type into Vizard Agent?
Describe the behaviour rather than the props when you brief Vizard Agent. "A student on her phone" returns somebody posing with a phone for a stock photographer; "a student checking her phone while her group carries on working around her" returns the situation the case study is actually about.
Prompt
Variants worth knowing:
- "A clip for each behaviour." The instruction that changes the search.
- "Rather than generic classroom shots." Rules out the obvious.
- "Finish on question cards." Makes it a teaching object.
What does Vizard Agent actually do?
Here is the Vizard Agent sequence on a real classroom case study built for discussion. Most of the work is sourcing footage, and the sourcing is organised beat by beat through the case rather than by subject matter.
- Reads the case and identifies its beats.
- Searches for a pupil distracted by a phone during class work.
- Searches for collaborative work between other students.
- Searches for comparison and social pressure as a visual idea.
- Searches for a supportive conversation between young people.
- Searches for reflection shots to sit behind the question cards.
- Downloads a candidate for each beat, including a phone being hidden.
- Builds one contact sheet of every clip downloaded.
- Inspects the frames together to pick the ones that cohere.
- Checks the supporting images work as backgrounds and cards.
Step eight is what stops the film looking like a stock reel. Each clip is plausible on its own; seen together, three different classrooms, four different age groups and two different countries are obvious, and the set has to be chosen as a set.
Step four is the one that needs a human idea first. "Comparison and social pressure" is not in the case study as a noun — it is the thing the case is about — and turning that into something filmable is the real editorial work.
Step six matters because the questions are the point. A card over a reflective shot invites a pause; a card over a black screen reads as the end of the video, and the class stops watching rather than starting to talk.
What does the result look like?
Three minutes that set up an argument rather than settling one. The behaviours are recognisable to the pupils watching, the class on screen reads as a single class, nothing in it identifies a real child, and Vizard Agent ends the film on questions rather than on a conclusion.
When does this not work well?
Cases that turn on something invisible are hard to film. Internal states, long time spans and anything that hinges on what a person is thinking will not be carried by stock footage, and narration or text has to do that work instead.
- Internal states. Nothing to photograph.
- Long timescales. Weeks do not show in three minutes.
- Very specific settings. A particular school or piece of equipment.
- Cases with a named outcome. Then it is a story, not a discussion.
- Sensitive material. Some cases should not be dramatised at all.
How do you fix a result that came back wrong?
Say which behaviour is not reading on screen. Vizard Agent assembled the film beat by beat through the case, so a note about the group-work section changes that section's footage and leaves every other beat exactly as it is.
- "That does not look like distraction." That beat is re-searched.
- "They look like different schools." The set is rechosen for coherence.
- "It is too obvious who is at fault." The framing is made more neutral.
- "The questions are leading." They are rewritten as open questions.
How does Vizard Agent compare to doing it yourself?
By hand you search for "classroom" and "student phone", get several hundred results that all look like advertising, and pick five that are pretty. The film looks professional and the class has nothing to discuss because nothing in it resembles their experience.
| By hand | Vizard Agent | |
|---|---|---|
| What is searched | The nouns in the case | The behaviours in it |
| The abstract beats | Skipped | Turned into filmable ideas |
| Choosing clips | One at a time | As a set, on one sheet |
| The ending | A conclusion | Question cards |
| Real pupils | Sometimes filmed | Never needed |
Common questions
Do I need to film real students? No, and you should not. Vizard Agent builds the case from stock.
How long should it be? Two or three minutes. Long enough to set up, short enough to leave time to talk.
Can it suggest the questions? Yes. Ask Vizard Agent for open questions rather than ones with an answer.
What if the case is abstract? Vizard Agent turns the abstract beats into filmable ideas, or says it cannot.
Will the clips look like one class? Only if they are chosen together, which is why the contact sheet matters.
Can it avoid taking sides? Yes. Say the framing should stay neutral and Vizard Agent keeps it open.
What about narration? Optional. Ask Vizard Agent for text cards and no voice; many cases work better that way.
Can it do several cases as a set? Yes, and keeping the same visual treatment across them helps.
What age group should the cast be? Tell Vizard Agent, because it changes every single search.
Can I use my own footage? Yes, if you have consent. Vizard Agent will mix it with stock.
Does music help? It can steer the judgement. Ask Vizard Agent for a neutral bed, or none at all.
What if the case names a real incident? Then anonymise it first. Vizard Agent works from what you give it.
Can it make a version for teachers? Yes — the same scene with the discussion notes added.
How do I stop it feeling like an advert? By searching behaviours. Advertising stock is what noun searches return.