Versely
    Guides

    ATS product video is the UI still, not cinematic office Veo

    I2V the dashboard screenshot after the UI is locked.

    Versely Team••4 min read

    An ATS is software. The product video is the UI. A Veo clip of a glass office with a laptop that almost looks like your dashboard is a brand film of a building you do not sell. Prospects want to see the pipeline and the scorecard. They cannot tell from bokeh.

    SaaS founders already know the asset that gets requested before a sales call: a 30-second product-in-use demo. Raw screen recording of the actual interface outperforms a polished mockup. Image-to-video does not change that test. It only adds motion to a frame you already own.

    The product is the screenshot

    Lock the UI. Then capture it. Then, if you want motion, run image-to-video on that still.

    Image-to-video is contracted to your frame. What plays at 0:00 is the file you sent. Column headers and the stage names you actually ship survive if they exist in the screenshot. A text-to-video prompt that says "a cinematic ATS dashboard in a sunlit office" invents chrome and a green button that is not in your app. Users notice when the click target they were promised does not exist.

    That failure is the SaaS rule: never generate your own product interface. I2V of a locked screenshot is the exception that still obeys it — the interface is photographed, not dreamed. Animating a still you already have is this job. Text-to-video is a different bill.

    Do this after the UI is locked. Animating last sprint's settings page is how you ship a demo of a control you moved. Treat the screenshot the way you treat a pack still: approve the pixels, then buy seconds.

    Clean account. Realistic-but-fake candidate names. No real PII in the frame. Whatever is on that monitor is what a security review will treat as in-scope.

    Motion prompts, not scene prompts

    The still is the product. The prompt is only the move: a slow push-in on the pipeline board, a gentle parallax, steam-free, no invented cursor ballet. Do not re-describe the ATS. The model can already see it.

    Short clips hold identity better. A long I2V is extra time for the model to drift the type, merge columns, and "improve" your data table. If you need a 60-second walkthrough of "how a recruiter moves a candidate from screen to offer," screen-record the flow. Use I2V for the hero loop on the landing page — the board that has to look alive without becoming a different product.

    A 9:16 upload animates 9:16 and a 16:9 upload stays wide. Capture the still in the ratio you will ship so there is no re-crop that eats the kanban. If the only copy of the UI is a small crop, upscale it first — an image-to-video model can only preserve detail that exists in the frame you gave it.

    Cinematic office is a different movie

    Veo is good at rooms, light, and people walking through glass. That is a founder-story plate or a "why we built this" cut. It is not the ATS. Mixing the two in one generate — "cinematic office, laptop shows our recruiting OS" — is how you pay cinematic rates for a hallucinated UI.

    Split the library:

    • Product. Screenshot, locked, I2V or screen capture.
    • World around the product. Generated b-roll, office, founder, if you want it.
    • Talking. A face and a script, not a fake monitor in the background of the face.

    The same demo clip should also live on the landing page and in sales collateral. A cinematic office loop does not answer "what does this actually do." The locked dashboard does.

    FAQ

    Can I prompt Veo to "show our dashboard" from a description?

    You will get a dashboard. It will not be yours. I2V the screenshot after the UI is locked, or screen-record the flow.

    When is the UI "locked" enough?

    When the nav, the primary action, and the board columns will not move before the clip has to live on the site. If a redesign is next week, wait. Do not buy motion on a frame you will retire.

    Should the demo be polished or raw?

    Raw, real interface. SaaS buyers want to see the genuine UI they would actually use. Motion on a real still is polish that does not replace the pixels.

    Can I I2V a Figma frame instead of production UI?

    Only if that frame is what ships. A marketing comp that does not match production is the same lie as a generated dashboard, just earlier in the pipeline.