- SAP has embraced generative AI, including using it to create art for digital billboards.
- Its CMO describes what she’s learned so far, and where she is most cautious about using the tech.
- She predicts generative AI will be a permanent part of her marketing strategy in four to six months.
The tech giant SAP has dived head-first into generative AI, and is in the midst of conducting about eight experiments to figure out how the new technology can power its marketing. It’s currently piloting how the technology can be used to make its ads better, make its marketing more personalized, and do a better job routing customer questions to the right departments.
The tech giant has already wrapped up and learned from other recent generative AI pilots, including a campaign in May when digital billboards in New York, Atlanta, and Los Angeles featured an image that changed every day, based on the biggest headlines in culture and business. The art was created by generative AI and touched up by a human artist.
“It ended up being an interesting experiment,” SAP’s chief marketing and solutions officer Julia White told Insider. The results were promising, with 3X the amount of impressions that billboard ads typically drive, she noted.
SAP largely sells to other businesses and professionals, so White said the next generative AI-powered campaign won’t focus on billboards. “The marketing investment mix as we move forward will be more social, digital, and direct, versus out-of-home,” she said.
SAP’s marketing team also used generative AI for tasks that are less splashy, more internally-oriented, but just as important. For instance, SAP involved 200 of its employees in a three-month experiment to test the tech on tasks like reaching out to prospective customers, creating content that would uniquely appeal to each prospect, and doing customer or industry research.
“We really saw efficiencies in that,” White said. “People will be able to get through more customer calls, more customer engagement, move customers through the customer journey faster. That’s something we’re going to continue.”
As much as White was impressed at how effective generative AI could be for certain tasks, she also learned of its limitations. In one exercise, SAP employees used generative AI to create a script for product demos, but the scripts occasionally referenced products that didn’t actually exist.
SAP also has to be very careful with the data they use to train generative AI tools, which are often built by different tech companies.
“Considering the data in our systems — financial data, HR data, supply chain data — this is incredibly sensitive and proprietary,” White said. “We do have the highest standards around responsible AI. We had explicit conversations with customers about whether they’re willing to share anonymized data to help us train, learn, and create benchmarks.”
With a handful of generative AI experiments wrapped up and more currently running, White is trying to figure out how to scale the tech across SAP’s marketing program, so it could become a permanent part of her strategy. Part of this consideration includes figuring out whether to continue partnering with generative AI solution providers or build the tools internally.
“I can see the promise in these pilots we’re running, but how do we do it in a systematic way? That’s the answer I don’t have right now,” she said. White anticipates that in four to six months, SAP will start to have generative AI figured out enough to make it a more permanent fixture of its marketing strategy.
Read the full article here