Salesforce launched Marketing GPT and Commerce GPT in a move that melds generative AI, Data Cloud and business outcomes.

At its Connections event, Salesforce outlined Marketing GPT and Commerce GPT. Generative AI has been a fixture on the earnings conference calls of both enterprise customers and vendors.

Enterprise vendors have been racing to embed generative AI into products while customers have been pondering strategy, compliance and tuning models with corporate data. Recent developments include:

Salesforce is looking to allay enterprise concerns about generative AI by connecting it to business outcomes and return on investment. If successful, Salesforce can also boost sales growth for its Marketing and Commerce unit, which grew 10% in the first quarter and was the company's slowest growing unit.  Here's the breakdown of Marketing GPT and Commerce GPT components:

  • Marketing GPT uses generative AI models and Data Cloud first party data to deliver personalized experiences.
  • According to Salesforce, Marketing GPT will enable marketers to create audience segments quickly with natural language prompts. In addition, Marketing GPT will auto-generate email content and personalized messages.
  • To boost marketing returns, Salesforce outlined Segment Intelligence for Data Cloud. The service connects first-party, revenue and paid-media data for a view of engagement. Marketing GPT will also automatically resolve customer identities and updates segments in Data Cloud.

Commerce GPT will enable customized commerce experiences and deliver auto-generated insights and recommendations.

These insights and generative AI tools can be aligned with business goals via Goals Based Commerce, which automates growth and conversion strategies. Goals Based Commerce is powered by Data Cloud, Einstein AI and Flow.

Constellation Research analyst Liz Miller said the Goals Based Commerce has the most potential to meld large language models and business outcomes.

Miller said:

"This is about bringing specific business goals like 'grow volume of purchase' or 'increase margin' and then applying AI to understand what is and is not working and the recommendations on how to optimize and improve. So, this isn't just an isolated point of analytics but rather bringing together multiple models, multiple points of data and multiple workflows to actually achieve better outcomes and allowing actual business goals define what better means."

Goals Based Commerce, which will be in pilot in October and generally available in February, can be configured to offer recommendations on strategies to improve margins and increase average order value.

Components of Commerce GPT include:

  • Dynamic Product Descriptions, which fills in missing catalog data and auto-generates personalized product descriptions.
  • Commerce Concierge, which engages customers with natural language interactions.

Early customers include Rossignol and YETI. Most of the Marketing GPT and Commerce GPT components will be in pilots in October with GA in February.

Salesforce is also going deeper with Google and enabling customers to bring their own models via Google Vertex so enterprises can centralize models from Vertex registered with Salesforce Data Cloud. Salesforce also has native integrations with AWS, Snowflake and Databricks. At Google I/O 2023, Google Cloud launches Duet AI, Vertex AI enhancements