Brands lose an estimated $130 billion a year on ads that simply don’t perform. For channels where creative does the heavy lifting, such as social, video and display, creative alone accounts for the majority of that waste — an estimated $35 billion.¹ Creative is the last major variable in performance marketing that remains largely unoptimised. Everything else, from buying to bidding to targeting, is now heavily automated. But turning performance insights into a new, brand-approved ad still takes time and every day that passes means more budget is spent on underperforming creative.
This lag exists because the signal and the response are disconnected. Performance flows through multiple people, platforms and approval processes before a new ad goes live. One team analyses results and decides what to change. Another interprets the plan, develops new creative and routes it through feedback and approvals.
But what if your next opportunity to improve ROAS surfaced automatically? What if the entire path from insight to activation happened without manual handoffs?
Adobe GenStudio for Performance Marketing just launched a new MCP server that makes this a reality. Now in Early Access, the MCP server lets marketers stay in the AI tools they already use, including ChatGPT and Claude. From there, teams can go from a question to a campaign refresh in a single conversation or automate tasks like performance reports that used to take weeks or even months to receive.
Imagine a system that never clocks out, catches the moment an ad is worth scaling or a signal is worth acting on and works toward your performance goals in the background — whether you're in a meeting, at the gym or asleep.
Conversational, always-on ad optimisation.
Performance marketers can now ask which ads are underperforming, understand why, explore recommendations conversationally and generate new creative variants — all without leaving their AI assistant of choice.
This represents a meaningfully different way to get an answer. A traditional dashboard can only answer the questions it was built for, with preset fields and filters. Through MCP, a marketer can ask nuanced, open-ended questions and get answers that are grounded in their own campaign performance, brand guidelines and historical results.
GenStudio for Performance Marketing continuously looks at all your live ads across channels to identify which creative attributes are driving results — from headline messaging and image style to offer strategy and emotional appeal. Based on this analysis, it proactively surfaces statistically significant recommendations on what to change and why, flagging winning ads to scale and underperforming ads to refresh. These recommendations are available directly in ChatGPT and Claude.
From there, teams can click to view creative options based on each recommendation and generate new ad sets on the spot. Without leaving the conversation, they can send those new ads for review, check approval status and activate them across channels.
That’s the conversational layer. Point the same MCP server at ChatGPT Tasks or Claude Cowork and those workflows become proactive instead of reactive. A performance marketer can set up a task that pulls all underperforming ads into a report and delivers it daily at 7am. Or they can create one that goes a step further, taking the top recommendations from GenStudio for Performance Marketing, generating fresh creative and dropping new drafts into their inbox — ready for review when the workday begins.
That same recommendation engine also powers a new experience in the GenStudio for Performance Marketing Insights tab. From any recommendation, you can click to generate your next best creative. GenStudio for Performance Marketing shows you multiple creative directions based on performance data, each representing a high-impact opportunity to lift performance and optimise in-market spend.
Pick one (or a few) directions to test and your new ad set will be prepopulated on the GenStudio for Performance Marketing canvas. Brand governance will be built in, with each ad checked against your brand, channel and accessibility guidelines. Review brand scores across the ad set, make any last-mile edits, then send the ads for review and approval before activating them across channels.
GenStudio for Performance Marketing MCP, recommendations and next-best creative generation are now in Early Access for ChatGPT and Claude, with support for additional AI assistants and agentic platforms coming soon.
Where this is headed.
GenStudio for Performance Marketing closes the gap between signal and response, so teams can optimise continuously instead of periodically. Early results from our own Adobe Global Marketing Organisation (GMO) have shown a 21% increase in conversions while reducing wasted spend.
These capabilities are the first step toward a more connected, governed marketing system — one that not only surfaces opportunities but also helps execute against them within the guardrails of your brand. The result is faster optimisation, greater connectivity across agentic platforms and more opportunities to drive growth.
Explore Adobe GenStudio for Performance Marketing or ask your Adobe team about Early Access to our latest features.
Recommended for you
