SAS embraces generative AI for marketing

Leading analytics and data management provider SAS will integrate generative AI capabilities with its Customer Intelligence 360 marketing solution. Customers will be able to choose which generative AI providers to integrate into their SAS instance, including popular models like ChatGPT, open source models and their own proprietary models.

Marketers will be able to use these capabilities to create marketing plans and build audience segments as well as generate personalized text content.

Specific capabilities. At the SAS Explore conference this week, the following core capabilities were announced:

• Integration via the SAS connector framework to large language model (LLM) providers.
• The ability to use natural language processing capabilities to identify relevant audience segments.
• Rapid content generation using a combination of LLMs and internal data sets.

Dig deeper: The rise of generative AI:  A marketer’s guide to textual and visual AI content

Use wise guardrails. SAS, like other major vendors, is emphasizing the importance of safety and responsibility to mitigate risks involved in generative AI use.

“SAS Customer Intelligence product development,” he said, “is tightly aligned to guidance from the SAS Data Ethics practice,” said director of digital marketing solutions Mark Chaves, speaking to media and analysts. He laid out a framework for governance:

“Prioritising data privacy. No sharing of company or customer data with the models.

“Maintaining human oversight. AI-generated content should always be reviewed and approved by humans.

“Creating interpretable and transparent output. It should always be clear to marketers how an AI algorithm arrived at its conclusions and recommendations.”

Mark Chaves, SAS

Why we care. Marketing Solutions is only one part of SAS Institute’s wide analytics and data practice that serves many functions within the enterprise. When it comes to introducing genAI for marketers, it’s no surprise SAS would be emphasizing the importance of guardrails — especially the third of the three pillars Chaves laid out above. Marketers in tightly regulated spaces like finance need to be confident that AI recommendations are not only sound but also possible to explicate.

The human-in-the-loop recommendation makes sense, of course, but it remains to be seen how that can scale to address content that can be generated in huge volume and at mind-warping speed.