Today at TrailblazerDX, Salesforce officially released Einstein 1 Studio, the backend to Copilot that will power the next generation of apps built on the Salesforce platform.
Einstein 1 Studio consists of the Prompt, Model, and Copilot Builder, and will enable Salesforce professionals to build templated AI solutions to roll out to their users. Salesforce has made many big bets over the last few years, but I feel this is the start of something huge.
What Has Been Released?
There’s been a lot happening in the world of Salesforce and AI over the past few months, so let’s break down exactly what Salesforce has released.
Last week, Salesforce announced the public beta of Salesforce Copilot, that is, the AI-powered chatbot, that will sit across all Salesforce products such as Sales, Service, Marketing, and Commerce Cloud. This new little widget could be compared to ChatGPT, but it sits inside your Salesforce org, and is grounded to answer questions about your own private company data.
Salesforce Copilot is mostly aimed at users of Salesforce, to help them answer queries about their customers, and perform certain actions. Whilst Copilot is cool, the scope of what it can perform would be nothing without the custom tools that Salesforce have just released. These tools could be compared to Salesforce Flow and will power the next generation of AI-enabled apps.
Salesforce Einstein 1 Studio
These tools come in the form of the Einstein 1 Studio, which includes Copilot, Prompt, and Model Builder. Let’s break down exactly what each of these tools do.
Copilot Builder, now released in beta, allows you to customize the user-facing Einstein Copilot chatbot. By utilizing a set of tools, you can integrate Apex, Flow, and MuleSoft APIs, and generative AI components to create custom functionality.
As Copilot is very much designed to be used whilst users are in the flow of work, just like talking to a digital colleague, custom actions can be created to help with time-consuming tasks. On one Salesforce demo, they asked Copilot to craft, and send an SMS using a Flow action.
Prompt Builder is perhaps the most exciting feature to come out of the Einstein 1 Studio release. This tool is now generally available (GA), and allows you to create templated, reusable prompts that can be deployed across your entire org in existing functionality such as Custom Buttons and Flow.
Prompt Builder essentially extends Salesforce’s existing low code suite of tools and allows you to integrate GenAI across your whole org. For example, a custom prompt can be seamlessly embedded in a contact record as a button, enabling an agent in the contact center to get a snapshot of all escalated cases for a customer in one click.
Model Builder allows you to choose the Large Language Model (LLM) for a specific use case, as opposed to other solutions, such as ChatGPT. Salesforce believes that this will become extremely important in the future, as specific AI models develop, and businesses build out their own LLMs.
Customers can use predictive and generative AI models from Salesforce partners, including Anthropic, Cohere, Databricks, Google Cloud’s Vertex AI, OpenAI, and train select models on Data Cloud data without moving or copying the information.
The AI Revolution Is Coming
We’ve been told for well over a year now that AI is coming, and it’s going to change everything. However, it’s sometimes hard to create the connection between these outlandish statements, and practically using AI in your everyday workflow.
With the Einstein 1 Studio, I think Salesforce has just given us a glimpse into what the future is going to look like. This revolution isn’t going to happen overnight, these tools will need to embed themselves within organizations, and most importantly, we will need to learn to be creative with them.
Salesforce has shown us some impressive demos with the Einstein 1 Studio, but the current use cases have generally been focusing on summarizing data, generating basic text for emails, as well as performing actions, such as generating an SMS message and sending it automatically. Salesforce has to start somewhere and get customers bought in, so I have no doubt this is the right strategy in order to lay the foundation for AI in the future.
But there are a few areas where AI has the potential to change everything within a Salesforce org, and potentially, the B2B world as a whole.
Firstly, with Data Cloud, Salesforce enables you to store vast amounts of data about your customers, building a complete custom profile. This can include in-app usage from your website, products purchased from the ERP, and marketing materials interacted with from your marketing automation system. As we all know, data is what powers AI, and the more context we can provide, the better and more accurate the responses will be.
Secondly, Model Builder gives you the ability to select different Large Language Models depending on the prompt you are creating. As most will know, ChatGPT has one of the largest, most generic LLMs, which isn’t trained for any specific purpose. This means the answers it provides, whilst impressive, are very generic.
The true power of AI will come when specific LLMs are created for specific purposes, e.g. create the ultimate LLM for negotiating a deal, or an LLM to provide great customer service and solve issues with a specific website or app.
“Our model is open and provides the flexibility to connect to a variety of different AI models. We’ve started to fine-tune models to suit unique employee or customer needs.” Clara Shih, CEO of Salesforce AI
Thirdly, I think the biggest impact to AI on businesses will come down to the Salesforce community, and the creativity that is exhibited by businesses using elaborate prompts, unique ideas, and exotic LLMs. This is fresh ground for all of us, but the fact that Salesforce has given us a tool much like Flow, which enables us to build anything that our heart desires, means that we should see some pretty amazing solutions built.
Summary
I would imagine Salesforce will be paying close attention to how the Salesforce community is using the Einstein 1 Studio over the next few months. Just as Salesforce Flow has been developed over the past few years to be the ultimate automation tool, I’m sure Salesforce will look to do exactly the same with its AI Studio.