When ChatGPT launched in late 2022, Doug Tallmadge, who was then an engineer at SpaceX, began to develop a deeper understanding of what AI was capable of. He envisioned an AI platform where all the different parts of a workflow would come together.
By early 2023, Gradial, an AI startup focused on marketing operations, was born, with Tallmadge at the helm as CEO and co-founder.
On Wednesday, Gradial launched new capabilities that allow it to go beyond revision and actually develop new, on-brand creative assets at scale.
Enterprise software systems have historically demanded a lot of cross-team communication, said Tallmadge. The marketing team might oversee brief writing and incorporating revenue goals, while the product team is building the actual product and designers are revising ad creative based on feedback from tests run by performance marketers.
Gradial was designed to bring together all the moving parts.
In addition to connecting teams, Gradial also connects back-end systems, like an “enterprise marketing harness,” Tallmadge said. (A harness, he clarified, is a software platform that serves as “the supply chain for AI” by pulling in all the necessary context for a given task like inputs and prompts.)
Gradial’s customers can unify all the disparate systems that their data lives in, from Jira to SharePoint to web channels and email platforms, through custom integrations. Those integrations create what Tallmadge described as a cohesive “system of record” that Gradial’s AI uses to develop content that not only fits brand guidelines but is already proven to perform well.
To each their own
Before it was able to launch fully developed assets, Gradial’s creative focus was on helping brands revise existing ads for different audiences.
For instance, a brand might need to have slightly different designs for ads in different languages. German words tend to be much longer than their English translations, said Tallmadge, so a design “that was beautiful in English looks really terrible in German.” Gradial’s agent is able to tweak the design, like shrinking or relocating the text, so it fits without looking cramped.
Marketers can also prompt Gradial to analyze creative performance and shift the budget from underperforming ads toward the top creative. The agent can “continuously analyze and remix” the content that’s being deployed on each channel, said Tallmadge, so no ad dollars are wasted on underperforming creative.
Of course, he added, a human always has the option to approve new creative before it’s launched. The amount of human oversight is up to the brand. Usually, Tallmadge said, marketers opt for human verification when making “big decisions” like whether to publish a new, AI-developed creative asset. But for changes like budget reallocations based on current performance, a lot of marketers are comfortable with a more hands-off approach. “It’s selecting from already approved content,” said Tallmadge, so “there’s no content risk.”
Something old, something new
But most of this is old news for Gradial. (Well, as “old” as any news can be for a company that popped up post-ChatGPT.)
What it couldn’t do in the past was develop new creative assets – at least, not reliably, and not at scale.
Rather than creating assets entirely from scratch, the latest version of Gradial favors something more akin to the Mr. Potato Head method – which is to say, assembling an entirely new ad from an assortment of existing assets.
Like marketers, Gradial has learned that it doesn’t need to “reinvent the wheel” every time it develops new creative, Tallmadge said; consistency, after all, helps build a cohesive brand identity.
Plus, if you just ask ChatGPT to generate an image for you, he added, it inherently “looks AI generated.” Gradial, on the other hand, looks at assets that already live in a marketer’s systems like Figma and SharePoint, and can combine approved backgrounds with well-performing copy, plus any approved logos and product imagery.
Gradial has been developing this capability in earnest for about three months, said Tallmadge, though it could have done the basics a year ago to make a single asset. But now, he said, Gradial can develop the 1,500 or so assets that a marketing team might need for a campaign, with different versions for channel, locality, various promo codes and so on.
The new development saves both time and money spent on AI. Gradial’s agents can develop hundreds or thousands of audience-specific creative variations using the same time and budget that previously would’ve only been able to target maybe a dozen audiences, said Tallmadge. Again, Mr. Potato Head: It’s all about the combinations and permutations.
Plus, Tallmadge said, Gradial charges its clients based on marketing output, rather than token use – a huge win for marketers, as token costs continue to increase.
The AI’s efficiency allows Gradial’s clients to “scale their vision,” said Tallmadge, “without scaling their dollars.”
