The Creative Brief Is the Design Now
What the Latest AI Design Wave Puts in Every Creative's Hands
By Keri Tomsic, PR and Content Lead for the Marketing AI Pulse community
The design brief is no longer just the starting point. For many teams, it is now the finished product. In the past few weeks, we have seen a rapid-fire launch of AI design tools: Canva AI 2.0, Claude Design, Adobe Firefly AI Assistant, and ChatGPT Images 2.0. Each one operates on the same principle: you describe the outcome, and the system delivers it. On May 20, 2026, Figma joined the race with its AI Design Agent in beta, cementing that this wave of innovation is accelerating, not slowing.
For marketers, this unlocks new creative leverage. For designers, it validates what they have always known: translating intent into structure is the most valuable skill in the room. The tools simply make that skill visible. Early results show that teams are learning fast, and the reality is more nuanced than the launch headlines suggested.
The Stack: What Launched and What Teams Are Finding
Canva AI 2.0
Canva's most significant launch in over a decade has turned the platform into a conversational, agentic workspace it calls a Creative Operating System. Brand Intelligence automatically applies your fonts, colors, and tone to every AI output. Connectors to Slack, Gmail, Zoom, and HubSpot pull in real work context. Practitioners who attended the launch called it a platform reset. The key takeaway: teams that invest time configuring Brand Intelligence see real workflow gains. Teams that skip this step get generic results. Learn more
Claude Design by Anthropic
Claude Design lets you describe a landing page, prototype, or pitch deck and instantly returns live, interactive HTML. It reads your brand's codebase and design files to apply your system automatically, then hands off directly to Canva or your dev team. Early adopters confirm it solves the brief-to-prototype bottleneck, which is where most creative workflows stall. One practical note: the Pro plan runs out faster than expected with regular use. Plan for higher usage tiers if you want to make this a core workflow tool. Learn more.
Adobe Firefly AI Assistant
Adobe's Firefly AI Assistant brings Creative Cloud tools into a single conversational interface. You describe the outcome, and the assistant sequences Photoshop, Premiere, Lightroom, Illustrator, and Express — no app-switching required. Frame.io integration lets stakeholders leave feedback directly, and the assistant applies revisions automatically. Firefly is built for commercial safety, with IP indemnification included. This is what gets teams through legal review. Learn more.
ChatGPT Images 2.0 by OpenAI
This is the first image model with reasoning built in. It plans layout before rendering, producing near-perfect text inside images. For marketing teams, this removes the single biggest production blocker. Teams are already using it for ad creative testing, localized campaigns, and LinkedIn carousels. The cost barrier for testing is now essentially gone. Learn more.
Figma Design Agent
Figma’s AI Design Agent, launched in beta, works right on the canvas and in the left rail. Teams can explore ideas, make bulk edits, apply design systems across files, and turn feedback into action without leaving Figma. Figma Buzz enables bulk, multi-channel asset creation from a single prompt. Figma Make connects designed assets directly to codebases. For product and UX-focused marketing teams, this finally closes the loop between design and development in a way other tools have not. The agent is rolling out gradually. Access is not yet universal, but worth activating now if your team spans design and code. Learn more.
The Skill That Connects All of It
The Designer Fund and Foundation Capital "AI in Design 2026" report, published this month and drawing on surveys of over 900 designers, puts data behind what practitioners are already experiencing: designers working with AI are doing more, owning more, and shipping more than a year ago. The tools have not replaced the judgment required to direct them. If anything, they have made that judgment more visible and more valuable.
Three Priorities That Will Define Your Team's Competitive Edge
Build the input layer before your team builds bad habits: The quality of AI output is only as good as the brand context you feed it. That means prompt libraries, pre-loaded design system components, and clear guardrails need to exist before anyone starts experimenting independently. Teams that skip this step get generic output and spend more time in revision than they saved in production. The creative lead's job is to build the input layer, not just review the output.
Recognize that the brief is now where creative leadership lives: Every tool in this wave executes on description. The strategic and creative thinking that used to happen across a week of concepting now must happen at the prompt level. The teams moving fastest are the ones whose leaders have gotten precise about articulating intent. That is not a junior skill. It is your skill, applied earlier in the process.
Solve one workflow problem completely before scaling: The instinct is to pilot everything. The smarter move is to identify the one place in your workflow where speed or consistency is costing you the most — asset resizing, localization, first-draft concepting — and use one of these tools to own that step entirely. A complete solution to a real problem compounds. A partial test of five tools does not.
The Gap That Still Needs You
Jasper's 2026 State of AI in Marketing report found that 91% of marketing teams now use AI, up from 63% last year. Only 41% can prove ROI. The tools are not the problem. Judgment, governance, and connecting creative output to business results still require a human. The teams seeing 2x and 3x returns restructured their workflows, built governance in, and stayed clear on which decisions need human judgment.
Strategic Takeaway
The brief-to-output gap has collapsed. Five major platforms launched in the span of weeks, each capable of turning a description into a deliverable. The question is no longer whether the tools can do it. The question is whether your team has the judgment to direct them, the governance to use them responsibly, and the workflow discipline to connect creative output to measurable results. The teams building that infrastructure now are the ones who will see 2x and 3x returns. The teams skipping it are generating output with no proven impact.
Bring your experiments. Bring your results. Bring the prompt that finally worked.
About Marketing AI Pulse
The marketers who figure this out first will have a significant edge. The Marketing AI Pulse community is where practitioners are already sharing what's working — real work, real results. Monthly briefing postcasts, live webinars, in-person sessions, and a peer network of marketers navigating the same shift.
Join free — the only cost is sitting this one out.
Frequently Asked Questions
-
Canva AI 2.0 turns the platform into a conversational creative workspace. You describe what you need, and it generates on-brand output by pulling from your fonts, colors, and tone automatically via Brand Intelligence. It also connects directly to Slack, Gmail, Zoom, and HubSpot so it can use real work context — not just generic prompts.
-
Claude Design is Anthropic’s AI tool that turns a text description into a live, interactive HTML prototype — a landing page, pitch deck, or UI layout — in seconds. It reads your brand’s codebase and design files to apply your visual system automatically, then hands the output directly to Canva or a dev team for production.
-
The Figma AI Design Agent, launched in beta on May 20, 2026, lets you describe a design outcome in plain language and generates it directly inside Figma. It is the first AI agent natively embedded in Figma’s design environment, removing the gap between creative brief and working design file.
-
Adobe Firefly AI Assistant is a conversational interface that sequences multiple Creative Cloud apps — Photoshop, Premiere, Lightroom, Illustrator, and Express — based on a single text instruction. Instead of switching between apps, you describe the output you want and the assistant executes the workflow. It includes IP indemnification for commercial use.
-
ChatGPT Images 2.0 is the first image model with reasoning built in. It plans the layout before rendering, which is why it can produce accurate text inside images — something earlier models consistently failed at. For marketing teams, this removes a major production blocker for ad creative, event graphics, and social content.
