Chapter 8What Is the Einstein Trust Layer, and Why Do You Need It?
Let me start this chapter by saying that typing Einstein Trust Layer over and over is suboptimal, and that the acronym ETL would be confusing for many of my readers, who are familiar with standard database extract-transform-load terminology. So let's just call it the Trust Layer.
We've already met the Trust Layer in the context of the Salesforce platform, and the purpose of this codicil is to round out the elements of this critical piece of technology. It's critical because it does nothing less than make LLMs actually useful to the enterprise, elevating them from novelty or black ops to something that can be trusted with your customer data and even your brand's hard-won reputation (see Figure 8.1).
The Trust Layer starts with a prompt. So it is not here to generate, optimize, or otherwise tweak the prompt. That's the job of the user, who may use Prompt Builder to construct and test prompts. Prompt Builder lets you build and test prompts that use merge fields and other data coming from sources like Data Cloud and Flows.
Once the prompt is created, it is passed to the Trust Layer – which, as we've said, mediates between the enterprise with its hard-working prompt engineers and the outside world of LLMs.
Figure 8.1 Einstein Trust Layer
Salesforce realized early on that it was too soon to pick a “winner” in ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access