Brand that travels with the work

On-brand work is not a filter you apply at the end. In Shyne, brand is part of the project: theme values, saved marks, and the files you have already said yes to.

Most teams already know their colors. They still ship a deck that looks like a different company because the colors lived in a slide master, a Figma library, a Canva brand kit, and a Notion page titled “please use these.” An agent that cannot see any of those will invent a fifth. An agent that can see a project theme has a chance to start where you already are.

We are not trying to become your system of record for every brand asset in the enterprise. We are trying to make the next sequence in this project look like the last one you approved.

What the teammate sees

When you ask for a new sequence, the agent can read the brand that is already there. It should not invent a second personality for every prompt. If a color or a logo is wrong, you correct it once on the canvas and it stays in the graph.

Every project carries a theme: shared colors, type, and spacing. Change the theme once and work built on it can follow. The file library holds uploads and generated media. The mark you attached on Tuesday is still the mark on Thursday. You do not need to paste it into the chat again unless you want a different mark.

Reusable brands can also be saved and listed through the product and the API. Use that when the same personality must start more than one project. A consultancy with three clients should not rely on prompt memory to keep those clients from blending. Save the brand. Open the right project. Let the teammate read what you saved.

Memory exists for the same reason. Things you want the teammate to remember — a phrase you always use, a phrase you never use, a product name with a precise spelling — should be saved, not hoped for.

One look, several outputs

A pitch deck, a LinkedIn carousel, and a launch film can share the same type and palette because they share the same project. Frames change size. The brand does not have to be rebuilt for each size.

This is the practical win. People do not fail at brand because they cannot pick a green. They fail because the green has to be re-entered every time the format changes. If formats are outputs and the canvas is the document, the green is already there when the format changes.

Export and publish compile that brand into the artifact. They do not open a second design file that slowly drifts. If the square crop is tight on the mark, you nudge the node on the canvas and inspect the rendition again. You do not rebuild the post in the scheduler’s editor, which does not know your theme.

Correct in place

If the teammate misses, select the node and fix it — or ask again with the frame still selected. The correction is the new source of truth. You should not have to maintain a slide master in one tool and a social template in another.

Correcting in place is also how brand stays honest. A library that is never applied is a moodboard. A theme that you actually change, on a frame people can see, is a decision. The next prompt should see the decision.

Be specific when the miss is systematic. “Stop using a second gold. Use the theme gold. Do not introduce a new sans.” That is a better prompt than “make it more on brand,” which invites a new personality with better manners.

What belongs in the project files

Put the official mark in the files, not in a chat attachment you will lose. Put the previous deck in the files if it is the one the client already approved. Put the product shots you are willing to publish in the files, and keep the shots you are not willing to publish out.

The teammate will use what it can reach. That is a feature when the library is curated. It is a problem when the library is a junk drawer. Treat project files like the brand kit you wish someone had maintained.

Small images and PDFs can ride with a prompt. Use that for the first session, then save the keepers. A launch week is the wrong time to discover that the only logo in context was a screenshot from 2023.

Theme is not chrome

Shyne’s own product chrome has a brand: green for interactive, gold for emphasis, Fraunces for display, Inter for interface. That is our UI. It is not your deck. Document themes are brand-neutral in the sense that they color end-user content, not our buttons. Your sequence should look like you. It should not inherit our pinwheel unless you put our pinwheel on it.

If you are making work about another company — a comparison, an integration, a case study — use their official marks, unmodified, and do not imply they endorsed the page. Nominative use is normal. Recoloring someone else’s logo is not. The teammate should be told that as clearly as you would tell a junior designer.

Brand across people and agents

A colleague who joins the project should not have to ask for the hex list. It is in the theme. A second agent, on a second seat, should not invent a new palette because it did not see the first agent’s chat. It should read the project.

That is the difference between brand-as-conversation and brand-as-state. Conversations evaporate. State stays. We will keep putting personality into state.

Roles still matter. A viewer can see the theme and cannot quietly replace it. An editor can make a bad color choice; you will see it on the canvas. Project creators and organization admins decide who else gets access. Publishing through an agent still pauses for explicit permission. Brand is not a reason to skip that decision. A beautiful off-brand post is still a post.

When the brand should change

Sometimes the next sequence should not look like the last one. A new campaign, a sub-brand, a confidential review with a different tone. Make that explicit. Duplicate the project, or change the theme on purpose, or create a sequence that carries scoped overrides. Do not hope the teammate will “feel” that this one is different.

Scoped variables exist so a frame or a sequence can diverge without lying about the project default. Use them when the divergence is real. Do not use them as a junk drawer of one-off hexes that should have been a theme change.

If you change the default, look at the sequences that were already built. Following the theme is not the same as rewriting history without you. Review the work. That is still a human job.

A first-hour brand setup

Create the project. Name it after the work, not after the tool. Set the theme to the three or four colors you actually use and the type you can defend. Upload the mark as SVG if you have it, PNG if you do not. Save a brand record if this personality will start another project. Attach last quarter’s approved deck. Then ask for the new sequence.

The first pass should already be closer than a blank prompt to an empty model. If it is not, the miss is useful. Fix the theme, not only the frame. The next pass will tell you whether you fixed a document or only a symptom.

Type, spacing, and the quiet parts

Color is the argument people enjoy. Type and spacing are the argument that actually shows. If the theme has a display face and a body face, say so in the first prompt. If the last approved deck used tight headlines and short lines, put that in memory. The teammate will otherwise give you a handsome paragraph that does not match anything you have shipped.

Spacing is a theme value, not a vibe. A sequence that suddenly has huge margins is usually a missed default, not a new campaign. Reset the frame to the sequence size before you invent a new layout system in a panic.

The quiet parts of brand are the words you refuse. Product names, capitalization, the phrase that sounds like a competitor. Save those. Correct them in place when they slip. Do not wait for a legal review to be the first person who notices.

What we will not pretend

We will not pretend a theme replaces a legal brand review. We will not pretend every official asset is in the library because a model has seen the website. We will not pretend “always on brand” is a guarantee. It is a default, a set of readable state, and a teammate that can be interrupted.

Brand that travels with the work is a boring capability on a good day. Boring is the goal. The exciting part should be the story on the canvas, not the weekly archaeology of where the logo went.