September 14, 2023

Spinnaker Support is proud to be here, in San Francisco, serving as your eyes and ears for DreamForce 2023. We just attended CEO Marc Benioff’s opening keynote session, during which he discussed Salesforce’s heavy focus on AI technology over the past five years or so.

Recently, all the buzz surrounding the release of ChatGPT and Large Language Models (LLM) has created a good deal of speculation as to how Salesforce would incorporate AI into their ecosystem. But, as the session drew to a close, it was clear to everyone in attendance that Salesforce is fully committed to heavily integrating LLM into every technology, while still honoring their organization’s core values.

For those who were unable to attend this session, we have a number of key takeaways for you:

Security Concerns of Using LLM’s

  • While LLMs, such as ChatGPT, have been garnering a ton of attention, when it comes to data security, most still have flaws— they generally absorb the data given to them and use it to improve the accuracy of their future responses. This lack of data security violates Salesforce’s core value of Trust.
  • Business technologists understand that AI’s rapid growth will impact a business’s ability to stay relevant, but they’re concerned about feeding sensitive data into this data-hungry technology.

Salesforce’s LLM Solution: Introducing the Einstein 1 Copilot

  • Einstein is a virtual assistant that’s capable of summarizing phone calls, creating action items, and making suggestions using a common language.
  • Einstein has a built-in “trust layer,” which scrutinizes all data input to ensure it doesn’t retain any sensitive information. Einstein’s trust layer is designed to be programmed by the admin, who explicitly configures which data can pass into the LLM and which data remains confidential.
  • Salesforce uses specific data from whatever record is in question (e.g. a lead or opportunity) and then uses Einstein to generate a response.
  • Einstein is also designed to assist internal employees in making decisions faster, so that they can spend less time on their keyboards and more time with their customers.

Upon reflection, we’re left with one lasting impression: We find it intriguing that so many new technologies are marketed for their ability to cut costs or do jobs faster by removing the human element from transactions, and that Einstein seems to do the exact opposite: Instead of humans needing to adapt to the format that a computer expects, Salesforce’s new LLM communicates in conversational language that even the least tech-savvy person can easily understand.

Stay tuned for more Dreamforce 2023 session coverage — and be sure to contact Spinnaker Support directly to chat about Salesforce’s newest innovations and how to apply them to your ecosystem.