How to Prepare Brand Assets for an AI Motion Project

A folder containing a logo, three screenshots, and a link to the website looks like a reasonable starting point for a motion project. Then the first draft arrives: the logo disappears on the dark background, the product name wraps in an unfamiliar font, and the interface screenshot still shows last month’s feature label. None of those problems began with the animation.

Preparing brand assets for motion design means making the supplied material usable in the situations the video will create. A logo needs an appropriate variant. A typeface needs the required characters and weights. A screenshot needs an identifiable product version and a clear role. The files matter, but so do the decisions attached to them.

A compact package should answer the questions the maker would otherwise have to guess: which asset is current, where it belongs, and what must stay intact when it moves.

Give each asset a job

Start with the intended scenes rather than a list of file formats. If the opening needs a product name, the middle shows an interface, and the ending asks the viewer to visit a website, those scenes create different asset requirements. You may need a wordmark, a clean screen recording, an approved URL, and a background-safe closing lockup. You may not need an illustration library at all.

For each supplied file, write a short role description. “Use this screenshot to show the completed booking state” is more useful than “Here are some screenshots.” The role tells the maker which information must survive cropping and which part of the image deserves emphasis. It also helps expose cases where the available image cannot support the planned scene.

Keep reference material separate from production material. A screenshot of a competitor’s video is a reference, not an asset you are authorizing someone to place in your own film. A low-resolution image from an old presentation might explain a composition without being suitable for final output. Mark that difference in the package itself so it does not depend on someone remembering a conversation.

A practical package has an obvious entry point. A short README can identify the approved files, explain what each one is for, and flag missing inputs. Avoid sending five versions of a logo without saying which version is current. The maker should be able to start with the right material without searching through an archive or guessing from filenames.

Supply the right logo variants and placement rules

A logo that works in a website header may not work against moving footage. Inspect it on the actual light and dark backgrounds you expect to use. Check whether fine details remain visible at the intended display size and whether a transparent file contains unexpected padding that makes alignment difficult.

Dropbox’s logo guidelines are a useful example of a brand treating its mark as a system with specific expressions, rather than as one universal image. Its published usage choices show what a loose logo file leaves unspecified.

For your own package, name the intended variant for each background condition. Include the minimum clear area if the brand specifies one, and explain whether the symbol can appear without the wordmark. If the mark must remain undistorted, say that explicitly. A broad instruction such as “animate the logo” leaves too much room to change the identity itself.

Supplying an SVG can preserve resolution, but it does not guarantee that every workflow will expose every path as an editable object. Supplying a transparent PNG does not make it unsuitable if its resolution is sufficient for the intended use. Choose the file with the actual composition and tool in mind, then check how it appears after import rather than treating the extension as proof of quality.

Do not bake an unnecessary background into the only available logo. If the end card changes color later, a white rectangle around the mark becomes an avoidable repair. Keep an approved original and let the composition provide the background where that is consistent with the brand rules.

The proof below uses the supplied OpenDesign lockup on two backgrounds. Its symbol, wordmark, and spacing stay the same; only the foreground treatment changes. This is a useful first asset check because a maker can see whether the whole lockup remains visible before adding a transition or product screen. Include a similar paired proof in your package when the film moves between light and dark scenes. Check the actual symbol-to-wordmark gap and the transparent padding around the exported file. Two files with different outer padding can appear misaligned even when both marks are centered mathematically.

Make the typography reproducible

A font name is an instruction, not a guarantee that the font will be available. The production environment needs access to the correct files or an approved loading method. It also needs the actual weights used in the design. A regular font made artificially bold can look different from the intended semibold, especially in large titles.

Dropbox’s typography guidance shows why type belongs in the asset discussion: its published system identifies particular typefaces and uses visible specimens. Your package should make your own hierarchy similarly concrete. A maker should not have to infer the heading face from a compressed website screenshot.

Test a sentence that resembles the video copy. Include the longest product name, the currency symbols you expect to display, and any accented characters. A typeface that handles a short English slogan may not contain the characters needed for a customer name or another market. Discovering that in a still proof is much easier than repairing it after every shot has been timed.

Provide a small set of text roles: main statement, supporting line, interface annotation, and qualifier. For each role, show an example and identify the intended weight. You do not have to specify every pixel before the design exists. You do need to communicate the relationship between the roles so that a legal qualifier is not mistaken for a second headline.

Keep usage permissions with the font source or internal asset record. Do not send a font file simply because it happens to be installed on one person’s laptop. If the team cannot distribute the file, explain how the authorized maker should obtain it, or agree on a permitted alternative before production begins.

The close view of Dropbox’s published mark makes another asset question visible: which geometric relationships define the identity? The diamonds share particular corners and gaps. Stretching the whole mark changes those relationships, even if the result still resembles a box. For your own logo, compare the imported mark with the approved original at a matched height. Look at shared edges, interior gaps, and small projecting details. Keep the reference beside the source so a later crop, mask, or nonuniform scale can be checked against the same geometry. A distorted mark is easier to catch in this isolated proof than amid a moving background.

Describe color as a relationship, not a bag of hex values

A list of colors does not explain how they work together. The maker needs to know which color belongs to the background, which carries the main text, and which indicates emphasis. Without those roles, every approved color can appear in the same frame while the result still feels unrelated to the brand.

Prepare one light composition and one dark composition if both are expected. Place the logo, a title, a supporting line, and a small annotation in each. This gives you a useful test of contrast, hierarchy, and the way the brand behaves at the scale of a video frame. It also reveals whether an accent color works only in large blocks rather than in small text.

Be especially careful with moving backgrounds. A title can be readable at the start and disappear when a bright object passes behind it. A still reference cannot establish readability across the whole shot. Decide whether the text needs a stable backing area, a quieter region of the image, or a different placement that remains usable as the scene changes.

If a gradient is part of the brand, describe its role and direction using an approved example. “Use the brand colors in a gradient” can produce an unlimited range of results. A reference that shows where the darkest area sits and how the text is placed makes the direction easier to reproduce without requiring the maker to guess your preferred blend.

Prepare product captures for their closest view

Choose screenshots according to the detail the video needs to show. If the key evidence is one status label, a wide dashboard capture may contain too much irrelevant material and too few pixels for that label. Capture the relevant state at a useful resolution and preserve an original before making editorial crops.

Record which product version or state the capture represents. A file can be visually clean and factually outdated. For a feature launch, the person responsible for the interface should confirm that the labels, controls, and result shown in the capture are appropriate for the intended release. Animation cannot correct a source image that represents the wrong workflow.

Use intentional demonstration data. Check names, emails, account identifiers, notifications, and browser tabs before recording. If you replace sensitive values, do it in the source environment when practical and confirm that the replacement does not change the meaning of the example. Do not hide the result the viewer needs in an effort to remove unrelated information.

Keep the language of the capture aligned with the article or video audience. An English explanation paired with a screenshot full of unfamiliar labels imposes extra reading work. If an English interface is available, use it. If the only valid source is in another language, acknowledge the situation and avoid pretending a translated overlay is an actual product label.

For a sequence, name the relationship between captures: before the action, during the action, and after completion. Do not assume three images are consecutive states just because they sit next to each other in a folder. The story should be supported by the actual product behavior those images represent.

Include motion behavior in the brand package

A brand can be consistent in color and type while moving in an entirely different voice. A restrained interface demonstration and an exuberant launch sequence may use the same assets but create different expectations. Describe the intended role of motion before choosing individual effects.

Dropbox’s motion guidance demonstrates that a brand system can include movement alongside static identity. Include behavior alongside static appearance in the package.

For a small project, translate broad tone into a few observable choices. For example, the product panel stays stable while an annotation appears; the title finishes entering before the secondary line begins; the closing logo holds still while the viewer reads the action. Use these relationships to judge whether the next scene belongs to the same system.

Explain what can vary. The opening may have more expressive movement than a pricing card. The feature result may deserve a brief emphasis while the disclaimer stays steady. A useful package gives different scenes a common voice without forcing them to perform the same animation regardless of their purpose.

If you attach a reference video, identify the part you want to borrow: the sequence of attention, the amount of movement, or the relationship between text and image. A bare link asks the maker to decide which of the reference’s many properties matters to you. That is a creative decision you can make more clearly before production.

Turn brand guidelines into a usable asset manifest

The manifest does not need to become another administrative burden. It is a compact index of decisions that could otherwise become expensive guesses. Keep it beside the files and update it when an approved asset changes.

A manifest should point to one approved version for each role. If you keep alternatives, explain why they exist. “Option B for dark backgrounds” is useful. “Final new latest 2” is not. The naming system should help another person make the same choice you would make.

Separate missing material into blockers and temporary choices. An unconfirmed price is a blocker for an offer shot. An undecided decorative background may be a temporary choice if the team agrees that it can change. This prevents production from stopping over a minor preference while a serious factual uncertainty slips through unnoticed.

Prove the package in one representative frame

Before requesting the whole animation, assemble a frame containing the elements most likely to conflict: the longest headline, the logo, a product crop, and any necessary qualifier. This stress test reveals whether the assets can coexist at the intended size.

Check the frame at the expected display scale. A composition that looks spacious on a large design canvas may become crowded in a social feed. Read the supporting copy without zooming. Look for logo padding, weak contrast, missing glyphs, and a crop that removes the important interface state. Fix the asset or rule that caused the issue rather than compensating with unrelated decoration.

Then ask for a short motion test where the background or layout changes. A static proof cannot show whether the logo remains readable throughout a transition or whether the text arrives too late to be understood. Keep the test focused on those uncertainties so that you can make a decision without commissioning a polished film first.

Supply asset roles with the files in your OpenDesign project, and check the imported logo and font in a representative frame before animating the rest. Keep that proof alongside the preview and export described in the Motion workflow.

Keep the approved package attached to the deliverable

When the video is approved, save the asset manifest that produced it. A later campaign update should not begin by guessing which logo, screenshot, or copy version was used. Keep the original captures and approved text available even if the export is the file people share most often.

If an asset changes after approval, identify which scenes depend on it. A new wordmark may affect only the ending. A new typeface can alter every line break and therefore the timing of several scenes. Treat those different impacts differently instead of automatically rerendering the whole project and hoping it still looks right.

The finished package should let someone answer three ordinary questions: which files are approved, how should they be used, and what must be checked if they change? When those answers are easy to find, the motion project starts with fewer guesses and remains easier to maintain after the first export.