AI Tools for Graphic Designers: Building a Full Campaign System in Stensyl.

Why a Single-Platform Workflow Matters for Campaign Work

Most graphic designers running campaign work aren't using one AI tool. They're using five. A separate image generator, a separate writing assistant, a separate research tab, a separate ad resizing tool — and somewhere in the middle of all that switching, the campaign's coherence quietly falls apart.

The problem isn't that any single tool is inadequate. It's that moving between them destroys context. The tone-of-voice rationale written in one tab never reaches the image generator in another. The competitor research done in a third tab never informs the copy hierarchy. Each output exists in isolation, and a collection of isolated outputs is not a campaign system.

Industry behaviour reflects this. Graphic designers are increasingly using AI across the full campaign arc, from brief writing and visual ideation through to social and ad delivery, rather than reaching for it only at the image-generation stage. The strongest use cases are text-heavy graphics, brand-consistent asset variants, and multi-format social production. That kind of workflow demands a connected environment, not five subscriptions in five browser tabs.

Stensyl's shared credit system is built for exactly this. Credits move fluidly across surfaces — Research, Write, Image, Graphics, Marketing Studio, Canvas, Boards — so a designer can move from competitive audit to final ad creative without paying per-tool premiums or losing the thread of the campaign. The brief written in Write can inform the image generated in Image, which gets collected into Boards, which feeds the Canvas pipeline that outputs to Marketing Studio. That's a campaign system. Not a collection of outputs.

What makes the difference at a practical level is the feedback loop between research, copy, visuals, and layout. When those four stages share a workspace and a credit pool, each stage can actually inform the next. Brand references travel. Tone-of-voice guidelines persist. Approved assets stay organised. The campaign has memory.

Stage One: Research and Brief Development in Ray and Research

Every campaign starts with questions, not assets. What visual language dominates the category? Which references are overused? Where is the white space? Before generating a single image, the Research surface gives graphic designers an audit of the competitive landscape that would otherwise take hours of manual tab-switching.

Stensyl's Research surface runs on Perplexity's backend, which means results are grounded in current indexed sources rather than training-data snapshots. For a packaging rebrand brief, for example, a designer might query the visual conventions of the premium food sector, surface which colour systems are saturated, and identify which typographic registers feel fresh versus clichéd — all before the creative direction is fixed. That audit becomes a documented artefact, not a set of mental notes that evaporate by the afternoon.

Once the research is gathered, Ray becomes the brief's first critical reader. Ray is Stensyl's AI assistant, running on full Claude Sonnet or Opus, with web search. It functions as a creative director you can argue with: feed it the research summary and ask it to challenge the brief's assumptions, identify which image generation models suit the campaign's visual register, or flag where the tone-of-voice is inconsistent with the target positioning. That kind of pressure-testing before generation saves significant rework downstream.

With the brief stress-tested, the Write studio takes over. The six-model picker lets designers select Claude Opus 4.8 for nuanced, brand-voice-consistent drafting, or GPT-5.5 for sharper, more direct copy structures. For a food packaging rebrand, this is where the full creative brief gets written: market positioning, visual references, tone-of-voice guidelines, typographic hierarchy, and the copy rationale that underpins every headline the campaign will eventually produce. All in one document, in one surface.

The critical step at this stage is saving everything to a Project workspace. Stensyl's Projects surface stores the brief, the brand identity, and the reference context so they're accessible from every downstream surface. The image generator knows what the brand sounds like. The Canvas pipeline knows what the brief requires. That shared context is what separates a campaign system from a sequence of unconnected generations.

"Moving from competitive audit to written brand rationale in under an hour isn't about speed. It's about keeping the thinking intact so it survives into the visual stages."

Practically: a food packaging rebrand might begin with a Research query on premium ambient food branding in Europe, produce a competitor audit in twenty minutes, move into Ray for brief critique, and land in Write for the full rationale document — all before lunch, all inside a single Project that remembers everything.

Stage Two: Visual Asset Generation Across Image and Graphics

With a brief saved to the Project, visual generation has something to work from. The choice of model matters here, and it's worth being deliberate rather than defaulting to whichever generator is most familiar.

For type-heavy hero graphics — the kind of work where the headline and the visual are inseparable — Luma Uni-1 is the current model of choice. It's built specifically for image generation with reasoning, web-grounded prompts, and strong typographic handling. That makes it well-suited to campaign work where the text treatment is doing as much compositional work as the image itself. For more photorealistic product shots, other models in Stensyl's library of 20-plus image generators will suit better depending on the reference direction established in the brief.

The Graphics studio handles the other half of campaign asset production: vector work, lockups, iconography, and pattern systems. A festival poster campaign, for instance, needs more than hero images — it needs a coherent graphic language that extends across merchandise, wayfinding, and digital formats. Graphics generates those elements from the same brief, keeping the visual system internally consistent rather than requiring a designer to rebuild style logic each time they switch between asset types.

Approved outputs from both Image and Graphics get collected into Boards. This is where the campaign's visual system becomes visible as a whole. Boards is a fluid canvas that merges reference collection and asset organisation: designers can group visual references alongside generated frames, compare iterations side by side, and see immediately where the campaign's colour system or typographic register is drifting. It replaces the practice of maintaining a separate moodboard folder and a separate output folder that never quite talk to each other.

For campaigns that extend into motion, Boards adds another practical capability. The first-frame and last-frame controls let graphic designers set precise start and end points for asset sequences, which matters when poster campaign elements need to extend into animated versions or short social video. The visual system established in still form doesn't have to be rebuilt from scratch when motion enters the brief.

The festival poster example is instructive. A campaign requiring eight poster variants across two colour ways, three layout orientations, and variable typographic weights can be generated iteratively in Image and Graphics, collected into Boards for comparison, and reviewed as a system before a single file leaves the platform. The iteration happens inside the workspace, not across a chain of export-review-export cycles.

Stage Three: Campaign Copy and Layout in Write and Canvas

Copy and visual generation are usually sequential in traditional workflows: write the headlines, then brief the designer, then generate the image. In Canvas, they can run in parallel.

The Write studio's six-model picker is the starting point for headline development. GPT-5.5 tends to produce punchy, short-form copy with strong rhythm — useful for OOH and social headlines where syllable count and cadence matter. Claude Sonnet 4.6 reads the brand brief more carefully and is better for brand-voice consistency checks: it'll flag when a headline drifts from the tone established in the rationale document, which is particularly useful when a campaign is generating a high volume of copy variants.

Running both models and comparing outputs directly in Write is faster than iterating on a single model. The differences become immediately readable: GPT-5.5 might produce something punchy but slightly off-brand, Claude Sonnet 4.6 something on-brand but lacking energy. The creative decision is then a genuine editorial choice rather than a guess about which model will perform better.

Canvas takes this further. The node-based workflow editor lets designers build pipelines that pipe LLM Chat nodes directly into Image Generate nodes, so copy variants and visual variants generate in parallel rather than sequentially. A practical pipeline for campaign work might look like this:

1. A single campaign headline enters the pipeline as the seed input.
2. An LlmChatNode branches it into three tone variants: confident and direct, warm and considered, and provocative and brief.
3. Each variant routes to a separate Image Generate node with a corresponding visual treatment drawn from the brand brief saved in Projects.
4. All three outputs collect into Boards for side-by-side review.

The RayNode sits above this as the agentic planner. For campaigns requiring multiple format outputs from a single brief — print, digital, OOH, and social — the RayNode can orchestrate the multi-step generation sequence, managing the order of operations and the routing logic so the designer is reviewing outputs rather than manually chaining prompts.

This is the part of the workflow that most clearly distinguishes a campaign system from a prompt-and-export approach. The pipeline is reusable. Change the headline seed, change the brand brief reference, and the same Canvas structure produces a new campaign's output set. The logic persists even when the content changes.

Stage Four: Social and Ad Creative in Marketing Studio

Marketing Studio is where the campaign meets its actual distribution formats. It handles social posts and performance ads under a single surface, using a Post and Ad framing toggle rather than requiring separate tools or separate workflows. For graphic designers who are responsible for campaign delivery as well as campaign creation, that distinction matters: there's no context switch between designing the organic social carousel and setting up the paid ad variant.

Carousels generate directly from the campaign's established visual system. Because the approved assets are already saved in the Project and collected in Boards, Marketing Studio can draw on them rather than requiring a designer to re-upload or re-describe the visual language. The colour system, the typographic hierarchy, and the image treatment that were established in Graphics and Image carry through. The carousel doesn't look like a different campaign.

Ad copy generation inside Marketing Studio works the same way. The research-backed copy feature aligns headline generation with the brief written in Write, so the ad headline isn't developed in isolation from the campaign rationale. If the brief says the tone should be direct and unsentimental, the ad copy reflects that, rather than defaulting to generic performance-ad language that has no relationship to the campaign's voice.

Reformatting for different placements is where campaign coherence most often breaks down in traditional workflows. Resizing a landscape poster for a square social post isn't just a crop — it requires rethinking the typographic hierarchy, the image composition, and the weight of graphic elements. Marketing Studio handles format adaptation within the same session, so a designer can move from a 1200×628px display ad to a 1080×1080px carousel frame to a 1080×1920px story without rebuilding the visual logic each time.

The practical example: a poster campaign for a product launch moves into paid social inside Marketing Studio. The hero image from the poster becomes the carousel's first frame. The headline, already written and brand-checked in Write, populates the ad copy field. The typographic system developed in Graphics informs the carousel's text treatment. The story format gets a recomposed version of the hero image, adapted for vertical viewing. All of this happens in one session, in one surface, drawing on assets that already exist in the Project.

Campaign Stage	Stensyl Surface	Primary Output
Competitive audit and research	Research + Ray	Brief rationale document
Creative brief and copy hierarchy	Write	Brand voice guidelines, headline set
Hero visuals and type-heavy graphics	Image + Graphics	Approved asset set in Boards
Parallel copy and visual generation	Canvas	Multi-variant output pipeline
Social posts and ad creative	Marketing Studio	Formatted campaign deliverables
Organisation and reuse	Projects + Boards	Reusable campaign template

Keeping the Campaign System Repeatable and Scalable

The difference between a finished campaign and a reusable campaign system is documentation. Specifically: what gets saved, where it lives, and whether it's structured to be handed off or reused without rebuilding from scratch.

Projects is the container for all of it. The creative brief, the brand identity settings, the approved assets in Boards, the Canvas pipeline — all of it lives inside the Project rather than scattered across a file system, a shared drive, and someone's desktop. When the next campaign starts, it starts from a template with memory, not a blank slate. The visual language is already documented. The tone-of-voice guidelines are already written. The Canvas pipeline already exists and needs only its inputs updating.

Credit planning across a campaign cycle requires some attention, particularly for studios running multiple campaigns in parallel. The Lite plan at £10 per month offers 1,000 credits and one concurrent generation — appropriate for individual freelancers working on single campaigns at a steady pace. The Starter plan at £22 per month raises that to 2,500 credits and two concurrent generations, which suits a designer who wants to run parallel copy and visual variants simultaneously. The Pro plan at £42 per month provides 6,000 credits and three concurrent generations, making it practical for campaign sprints where several asset families are generating at once. The Studio plan at £84 per month offers 12,500 credits and four concurrent generations for studios managing multiple client campaigns in the same billing period.

Concurrency is the practical constraint worth planning around. A Canvas pipeline that branches into three simultaneous image generations will queue at the concurrency limit of the active plan. On Starter, two of the three will run in parallel and the third will wait. On Pro, all three run simultaneously. For deadline-sensitive campaign production, that difference is worth factoring into the plan choice before the sprint begins.

Onboarding a second designer or a client reviewer into an existing Project doesn't require rebuilding context. The Project workspace is shared, so the incoming collaborator sees the brief, the approved assets in Boards, and the Canvas pipeline without needing a separate briefing session. Client reviews can happen inside the same environment where the work was made, rather than requiring an export-to-PDF cycle that loses interactivity and context.

What to canonise inside a Project for maximum reuse: the creative brief as a Write document, the brand identity settings including colour references and typographic guidelines, the approved hero assets collected and labelled in Boards, and the Canvas pipeline saved in its generic form with placeholder inputs rather than campaign-specific ones. That structure means duplicating the Project for the next campaign quarter is a matter of swapping the brief, updating the copy seed, and refreshing the seasonal imagery. The system's logic remains intact.

The Q3-to-Q4 transition is a useful test case. A campaign built in Q3 with a summer visual register — warm light, high contrast, saturated colour — can be adapted for Q4 by updating the brief's seasonal direction in Write, regenerating the hero assets in Image with an updated reference, and running the Canvas pipeline again with the new inputs. The pipeline architecture, the brand identity, and the copy hierarchy don't need to be rebuilt. They've already been established. What changes is the content. What stays is the system.

"The goal isn't to finish a campaign faster. It's to build a campaign system that makes every subsequent campaign faster, more consistent, and less dependent on starting from scratch."

Graphic designers who treat Stensyl as a campaign system rather than an image generator will find that the real efficiency gain isn't in any single surface. It's in the accumulated context: the brief that informs the image, the image that informs the ad, the ad that informs the next brief. Each campaign makes the next one easier to build and harder to get wrong.