Already at Dreamforce 2023, Salesforce has announced Einstein 1, the next generation of its Salesforce Platform.
The enterprise tech giant has now followed this up by launching Einstein 1 Copilot.
Einstein 1 Copilot is a GenAI-fueled conversational assistant, soon to be available in every Salesforce application.
As employees use those apps, they can engage with Copilot to complete various workflows and ask questions in natural language.
The answers come from company insights stored in Salesforce Data Cloud.
Introducing Einstein 1 Copilot in a media briefing, Clara Shih, CEO of Salesforce AI, stated:
Our customers have told us: “We love having these generative AI features in the flow of work, we love that they’re turnkey, and how innovations like Chat Reply Recommendations just show up.” But our customers have also told us they want more. That is why we are releasing Einstein 1 Copilot.
“Einstein 1 Copilot – a conversational AI assistant for both a company’s employees and their customers to safely and securely access generative AI. To do their jobs better, faster, and more easily.”
Copilot also incorporates the Einstein GPT features Salesforce announced in March.
For instance, Copilot may recommend a post-sales call action plan in Sales Cloud or auto-generate a new contact center knowledge article in Service Cloud.
Moreover, as Data Cloud is a foundational part of the Einstein 1 Platform – connecting all customer data – CoPilot can tap data from all Salesforce apps.
While it does so, the solution may generate more accurate AI recommendations and content.
That explains the tagline: “Everyone’s an Einstein with Copilot.”
As that suggests, Copilot will become available to every Salesforce user across all of its clouds.
In addition, Shih notes that Copilot’s capabilities will expand even further beyond the innovations announced in March.
“These were specific features,” concluded the CEO. “Now, it’s an open-ended assistant that employees and customers can simply chat with.”
Copilot Studio to Customize the AI Assistant
Alongside Copilot, Salesforce has announced Einstein Copilot Studio, allowing enterprises to customize the AI assistant to their business requirements.
In doing so, Salesforce further differentiates Copilot and sticks by its traditional platform approach to building software.
“We’ve always taken a platform approach to everything we build because we’ve developed an incredible ecosystem,” added Shih.
Customers want to make things their own, and partners have tremendous ideas and creativity that they want to bring onto the platform… Einstein Copilot Studio allows them to do just that.
The Prompt Builder and Skills Builder within the Einstein Copilot Studio are two of the most defining features that enable such customization.
First up, the Prompt Builder allows businesses to tailor the reusable prompt templates included in Einstein GPT – and now Copilot.
Employees can use these templates to more quickly engage with Copilot when asking a common question or performing a standard task.
Copilot then follows a pre-set course to resolution, drinking from customer data stored across the Einstein 1 platform.
Moreover, with the Prompt Builder, businesses may have Copilot reference specific database fields in Data Cloud when performing a particular task.
From there, the business may run A/B tests to roll out the best version.
Next is Skills Builder, which allows businesses to teach Copilot to perform new actions for them.
The tool may prove particularly valuable to businesses with some of their customer workflows already on Salesforce.
Why? Because with Skills Builder, those companies may select those workflows and familiarize Copilot with them.
Such a workflow may include resetting a password or initiating a return in retail. Giving this example, Shih stated:
With one click, that business can make it possible for employees in the customer service department – or even customers themselves – to access that workflow, not through clicks, but through natural conversation.
Finally, the Copilot Studio enables businesses to let the AI assistant loose on digital customer-facing channels, such as live chat, SMS, and WhatsApp.
Some may think twice about such use cases, worrying about the potential security implications. Yet, Salesforce aims to allay those fears with the Trust Layer that runs across the Einstein 1 platform – including Copilot.
A “Trust Layer” Secures Copilot
Most will recognize the promise and potential of Einstein 1 Copilot to drive growth, cut costs, and fundamentally improve customer service and customer experience.
However, some don’t trust the safety and security of generative AI. According to Salesforce’s own data, 52 percent of consumers do not.
There are many significant reasons for this, including the risk of LLMs taking their data, headlines of hallucinations, and reports of bias and toxicity.
Salesforce has put forward its tenets of trusted, ethical, and humane AI to help ease those fears. These are:
- Your data isn’t our product
- You control access to your data
- We prioritize accurate, verifiable results
- Our product policies protect human rights
- We advance responsible AI globally
- Transparency build trust
That forms the basis of the Einstein Trust Layer – as outlined below.
There are many notable features here. For instance, there’s data masking on anything sensitive and proprietary. Then, there’s a Zero retention SLA feature, interacting with in-house and external LLMS, ensuring that the model retains none of the context shared in the prompt.
There are also further features – including prompt defense and citation – not included in the graphic above.
Yet, this layer will evolve as Einstein 1 Copilot enters the market.
Indeed, Shih promised that Salesforce will keep collaborating with customers and partners, “red teaming” to ensure this stays the “most secure and robust trust layer.”
Eager to keep closer tabs on Salesforce’s ongoing innovation? If so, these other recent stories may also catch the eye:
- Genesys and Salesforce Announce a Joint Platform for CCaaS and CRM
- Salesforce on the Potential of Further Layoffs, GenAI In Customer Service, & More
- All the Highlights from Salesforce’s Winter ‘24 Release