Generative AI isn’t just for creating unsettling art, chatting with robots or, uh, attempting to break up marriages.
Soon, marketers will be able to use generative AI to train models that create audience segments and simulate customer journeys.
On Tuesday, Adobe announced plans to weave generative AI tools into its real-time customer data platform through Sensei, its AI and machine learning technology.
Adobe will allow marketers to train AI models with their own data and content within its CDP so the output is aligned with their brand style, said Ryan Fleisch, director of product marketing for the Adobe CDP.
Training day
Adobe released the first version of Sensei in 2016. It has since been integrated into multiple Adobe apps and services, including Creative Cloud, Document Cloud and across the Adobe Experience Cloud, which includes Adobe’s CDP.
In Photoshop, for example, designers can already use so-called neural filters to automatically edit and enhance images. In Acrobat, Adobe’s PDF software, there’s a tool called Liquid Mode that uses AI to automatically adjust PDF documents to fit a mobile screen.
Although generative AI has recently acquired a reputation for going rogue (or in some cases being flat-out wrong), allowing brands to train using their own data should help ensure quality and accuracy, Fleisch said.
Customer data platforms are a repository where brands store their data (typically first-party data) from multiple sources to create a single customer view.
But brands will also be able to choose whatever large language model (LLM) they want to use as their training data set. (An LLM is a category of machine learning model that trains on massive amounts of text to generate an output.)
Just ask
AdExchanger Daily
Get our editors’ roundup delivered to your inbox every weekday.
Daily Roundup
Adobe’s Sensei GenAI will rely on “multiple large language models for text-based use cases,” Fleisch explained.
Translation: Marketers will be able to use natural language prompts to generate their segments and insights, not unlike having a dialogue with ChatGPT.
A retailer, for example, could use natural language to generate a segment of customers most likely to respond to a promotion for an upcoming holiday campaign. Or a marketer in the travel sector could produce a segment of people most likely to get value out of a newly launched mobile app.
Or a brand might just want to play around and see what happens.
Marketers will also be able to use the tool “to find relevant segmentation that they may not have previously considered,” Fleisch said.
And for more involved projects, Adobe plans to release a feature called “generative playbooks,” which draws on historical data to give brands a foundation for generating new ideas.
For instance, say a group of customers recently had a subpar experience with a brand. The marketing team could use generative AI to help map out the best set of touch points most likely to change their outlook and win them back, like serving a specific targeted promotional offer.
But these capabilities aren’t live in Adobe’s CDP just yet.
According to Fleisch, Adobe plans to share more details about the availability of Sensei GenAI for its real-time CDP “in the months ahead.”
