B-roll is supplementary footage used to support the primary content of a video. It shows a product, a process, a location, an example or a contextual detail while the main narration or story continues. The primary footage is usually called A-roll: it carries the main message, and B-roll helps the viewer see, understand or feel more of it.
One nuance saves a lot of confusion: B-roll is defined by its role in the edit, not by the camera, file type or source used to create it.
A-roll versus B-roll
A-roll carries the primary message or action. B-roll explains, demonstrates, proves or visually supports it. A founder explaining why onboarding was rebuilt is A-roll; the screen recording of the shorter onboarding flow that appears mid-sentence is B-roll.
The same file can play either role. A screen recording inside a founder story supports the narration, so it is B-roll. The same recording inside a software tutorial carries the lesson itself, so there it is A-roll. Its function inside the final edit determines its role. And B-roll is not inherently less important; in the seconds it is on screen, it is the video.
What is B-roll for?
Good B-roll performs at least one of six functions:
- Explain: a diagram while an abstract idea is described.
- Demonstrate: the actual feature while it is being claimed.
- Provide evidence: verified, real material that shows the stated result.
- Add context: the place, the object, the situation being discussed.
- Support continuity: bridging an edit so a cut does not distract.
- Create emphasis: a close-up at the moment that matters most.
B-roll does not need to perform every function. It should perform at least one, and if you cannot name which, the clip should go.
The common types
Original recorded footage is the most specific and the most credible; its cost is that you have to shoot it. Screen recordings are ideal for software, with one standing review: check for private information, outdated interface elements, stray notifications and customer data before they go in. Screenshots and photographs work as stills; movement is not required. Stock footage is fast and broad, and its weakness is often specificity: the generic office is nobody's office. Diagrams and graphics explain processes better than any filmed scene (Upload, Processing, Edit, Review as four labeled steps beats forty seconds of screen time). AI-generated visuals can illustrate concepts that are hard to film, and they carry their own review burden: visual errors, incorrect text, and scenes that appear real but never occurred.
That last point is the one hard boundary in this subject: AI-generated footage should not be presented as evidence of a real event, person, customer or result. Looks realistic is not the same as documents reality.
Cutaways, establishing shots and inserts
Three terms you will meet everywhere. A cutaway temporarily moves away from the main subject to show a related detail, like the camera being discussed. An establishing shot shows the broader setting of a scene. An insert shot shows a specific detail up close. All three are B-roll jobs with different aims: context, place, precision.
Does every video need B-roll?
No. A direct talking-head shot can be the strongest choice, especially when facial expression, authority or personal accountability carries the moment. A video does not become incomplete because the same person remains visible for several seconds.
B-roll improves a video when the connection is specific: the words "our dashboard shows this in one view" deserve the dashboard on screen. It weakens a video when it is loosely related, hides the speaker's delivery at the wrong moment, competes with the captions, or implies evidence that does not exist. Every layer you add (speech, captions, B-roll, graphics, music) draws from the same attention budget; when the budget is spent, comprehension pays.
A quick decision test
Before any clip goes in: which exact sentence does this support, what function does it perform, is it factually appropriate, am I allowed to use it, does it fit visually, and is it on screen long enough to be understood? Six yeses, or no clip.
Where ReadyForm fits
ReadyForm places supporting visuals while it makes the complete edit from your own takes: it finds stock footage for the sentence it supports and uses the media you upload to your Library, timed to the words. Every placement stays visible and adjustable per scene, because the review questions above stay yours. See how the edit is made or the AI B-roll feature.