Flow AI for designers
Flow makes video, not stills, so the useful question for a designer is not how it fits your image workflow but where motion belongs in your process at all. Four workflows where it earns its place, and an honest account of where it does not.
Flowain · Last reviewed 2026-09-03
Flow does stills and motion now, which changes the pitch
Whisk and ImageFX were folded into Flow in February 2026, so image work and video work now happen in one place.
Flow generates video with synchronised audio, and since February 2026 it generates and edits stills as well, after Google folded Whisk and ImageFX into it. Older advice telling you to choose Whisk for images and Flow for video is out of date; both now live in the same tool, on the Nano Banana models.
What Flow offers a designer is access to motion without a motion department. The skills that transfer are the ones you already have: composition, light, and knowing what a frame is supposed to communicate. The skill you are missing is the vocabulary of camera work, and that is a short gap to close.
It is worth being blunt about cost before the workflows. Flow runs on Google AI plan credits at roughly 20 per Fast generation, and retries cost the same as first attempts. A designer exploring casually on the entry plan will run out in a week.
Four places Flow earns its keep
Ordered from strongest fit to most situational.
Previsualising a scene before anyone shoots
The strongest fit, and the one that saves the most money.
Blocking a sequence in Flow costs credits rather than a location, a crew, and a shooting day. For a director or a design lead, that changes what is worth exploring: you can test three approaches to an opening shot and discard two without anyone having carried a light stand.
Work shot by shot. Describe one camera move, generate eight seconds, and judge whether the framing communicates what you intended. What you are testing at this stage is composition and rhythm, not finish, so a Fast pass is the right quality setting throughout.
What you end up with. A rough sequence a cinematographer or client can react to, produced in an afternoon instead of a pre production week.
Putting an existing concept into motion
Frames to Video is the bridge from static design to moving image.
If you have already designed a key frame, you are holding the most valuable input Flow accepts. Supply your artwork as the first frame and a second image showing where the composition ends up, and Flow generates the movement between them.
Be clear eyed about what this is. Flow reinterprets your frames rather than preserving them pixel for pixel, so it is not a substitute for animating your own artwork properly. What it gives you cheaply is a sense of whether an idea works in motion at all, which is usually the question you actually need answered first.
What you end up with. A motion test that tells you whether a static concept holds up moving, before you commit to animating it properly.
Holding a look across a sequence
Saved ingredients are the closest thing Flow has to art direction.
The failure mode of AI video in design work is that every clip looks like it came from a different project. Different light, different palette, different face. That is what happens when each shot is written as a fresh description.
Save the character, the product, and any recurring object as ingredients, then reference them with @name in every prompt. Google calls this the most reliable way to carry the same subject from shot to shot, and it is what lets a sequence read as one piece rather than a collection.
What you end up with. A run of shots that share a look well enough to cut together without grading them into agreement afterwards.
Pitching with footage instead of description
A moving reference argues better than a static board.
A mood board asks a client to imagine the motion. Eight seconds of generated footage shows them. For pitch work that difference is worth more than the polish gap between generated and finished material, because the decision being made is about direction rather than execution.
Label it clearly as previsualisation when you present it. Passing generated footage off as finished work sets an expectation about the final deliverable that you will then have to walk back, and that conversation is worse than the one you avoided.
What you end up with. A pitch that communicates intent in the medium the final work will exist in, rather than a proxy for it.
Where Flow sits in a design stack
Before your existing tools, never instead of them.
Figma
Design your key frames in Figma, export them, and feed them to Frames to Video as the endpoints. Flow generates the movement; Figma keeps owning the composition.
After Effects or Premiere
Flow produces clips and stops. Assembly, typography, grading, and any text on screen all happen downstream. Scenebuilder will order and trim your clips, but it has no tracks or keyframes, and in-frame text is unreliable.
Photoshop or Procreate
Useful for preparing the first and last frames when you want a specific look that a text prompt will not reliably produce. Hand drawn endpoints work fine as inputs.
What will bite you on a deadline
Worth knowing before you promise a client something Flow cannot deliver.
Brand accuracy is the big one. Flow will not reproduce a logo, a specific product, or a named person correctly, and no amount of prompting fixes it. Anything requiring the real thing needs the real thing.
Typography in frame is unreliable enough that you should assume it will not work. Plan to add every piece of text downstream, and compose your shots leaving room for it.
Reproducibility does not exist across generations. If a client asks for the same shot again with one small change, you cannot simply regenerate it, and that is a genuinely awkward conversation to have mid project. Extend within a project rather than planning to recreate anything later.
Budget the credits before you quote. Iteration is the whole method, and every iteration bills. A project that needs twenty usable shots is a lot more than twenty generations.
The full capability and limits list is on the Flow AI features page, and prompt structures are on Flow AI examples. The end-to-end process, from a blank project to an assembled sequence, is on the how to use Flow AI page, and what Flow AI is covers the plans and credit costs behind the numbers quoted here .
Questions designers ask about Flow AI
The practical ones that come up when adding it to a process.
Do I need filmmaking experience to use Flow as a designer?+
You need the vocabulary more than the experience. Veo responds to shot size, camera movement, lens behaviour, and light direction. If you already think in composition and light as a designer, you are most of the way there; what you are missing is the naming, and that takes an afternoon rather than a career.
Can Flow animate a design I already made?+
Through Frames to Video, yes, in a limited sense. You supply your artwork as the first frame and a second frame showing the end state, and Flow generates the motion between them. Since February 2026 you can also generate and edit the stills inside Flow on Nano Banana rather than importing them. It is not rigging or true animation, and it reinterprets rather than preserves your artwork exactly, so treat it as motion exploration rather than asset production.
Is Flow output good enough for client delivery?+
For pitch and previsualisation, usually yes. For final delivery it depends entirely on the client and the category, because the output is generated footage with the failure modes that implies. Most designers using it seriously treat it as a tool for showing intent, not for shipping the final asset.
How does this fit with Figma or After Effects?+
Flow sits before them. It generates stills and footage and will assemble a straight sequence in Scenebuilder; your existing tools handle typography, grading, and any real editing. Nothing about Flow replaces a design tool, and expecting it to is the fastest route to disappointment with it.
What does this cost a designer per month?+
Nothing, to start. Every account gets 50 credits a day free with the full feature set, which is two Veo Fast generations, five Lite ones, or around a dozen short Omni Flash clips. For client work that runs out quickly, and paid plans add monthly credits on top: Plus at $4.99 adds 200, Pro at $19.99 adds 1,000. Most designers working to deadlines end up on Pro or higher.
Flowain
Writer and editor, FlowAIFX
Flowain writes and maintains FlowAIFX
Every page is checked against Google's published specifications for Flow and Veo 3.1 on the date shown, and corrected when Google changes them. Where a limit is undocumented or a credit cost is easy to misjudge, it is written down rather than left out.
Last reviewed 2026-09-03 · Report a correction
Start with a frame you already designed
Take a key frame from a project you have finished, make a second frame showing where it ends up, and run it through Frames to Video. It is the shortest path from what you already do to what Flow adds.