Getty Images/iStockphoto

Salesforce unveils Einstein 1 to embed AI in every app

The CRM vendor unveils a suite of new capabilities for its revamped AI platform, including a genAI Copilot in every app, and a way to customize the Copilot's functions.

Salesforce is making generative AI an integral part of its portfolio.

The CRM vendor on Tuesday unveiled Einstein 1, its revamped Einstein platform. The offering includes Einstein Copilot, Einstein Copilot Builder and Einstein Trust Layer, in conjunction with Salesforce Data Cloud to unify customer data and improve AI.

Salesforce unveiled Einstein 1 this week to audiences in San Francisco at Dreamforce, the vendor's annual user conference.

Einstein 1 will be available in pilot in the fall, according to Salesforce.

Copilot rivalry

Einstein 1 brings Einstein Copilot, Salesforce's generative AI assistant, into all Salesforce applications. Users will be able to work with generative AI for projects in sales, service, marketing and commerce.

It's not surprising that Salesforce and others have landed on Copilot as the label for their AI-enabled assistants. Key to adoption of AI is trust, and part of that means workers trusting that AI won't replace them, so Copilot strikes the right balance.
Rebecca WettemannCEO, Valoir

Salesforce isn't the first vendor to dub its generative AI assistant copilot -- Microsoft Copilot was unveiled in March. Valoir analyst Rebecca Wettemann says the name is appropriate, however.

"It's not surprising that Salesforce and others have landed on Copilot as the label for their AI-enabled assistants," Wettemann said, "Key to adoption of AI is trust, and part of that means workers trusting that AI won't replace them, so Copilot strikes the right balance."

Salesforce competitors, including Google, also have useful generative AI tools for CRM and contact center search resources, as well as enterprise productivity, according to Opus Research analyst Dan Miller.

"Both Microsoft, with its Copilot and attendant tools bolstering its Digital Contact Center Platform, and Google, with its Contact Center AI offerings and Vertex AI, present formidable, direct competition," Miller said.

Giving teams more control over AI output

Einstein Copilot's capabilities are strengthened by Einstein Copilot Studio, which lets users customize Copilot functionalities.

Einstein Copilot Studio includes a Prompt Builder, which helps users create reusable templates that are connected to company data. The Prompt Builder lets the user type in a prompt name and description to better define the task.

The Studio is an effective tool that gives users more control over AI output, according to Miller.

"[Einstein Copilot] is made more bulletproof by Einstein Copilot Studio, which gives subject-matter experts, rather than the models themselves, the ability to customize answers so that they conform to company policy, regulatory structures and common sense," Miller said.

Einstein Copilot Studio also includes a Skills Builder, where users can teach Einstein Copilot to perform new actions, and a Model Builder to let companies choose which generative AI models they want to integrate, without displacing or replicating data in Data Cloud.

AI accessibility and LLM inclusivity

There are several benefits to providing ready-made Copilots in every app and enabling use of multiple generative AI models. For one, it gives more employees the opportunity to use AI in their work processes.

"Salesforce offers options for customers with varying degrees of AI literacy and expertise to take advantage of Einstein's AI," Wettemann said.

It also gives various teams a common tool for self-service.

"Einstein Copilot provides a ubiquitous mechanism for every employee to do natural language searches of domain-specific and company-specific knowledge bases," Miller said.

In addition, it lets companies try different models.

"Having multiple LLMs -- I think it's a tremendous advantage," Salesforce AI CEO Clara Shih said during a press briefing.

"It's a very smart idea for companies to not lock into one vendor and actually take an open approach so that as the models evolve, we can easily swap in and out and optimize the right model for the right task," she added.

Mary Reines is a news writer covering customer experience and unified communications

Next Steps

Solo Salesforce acquisition in 2023 closes year of austerity

Salesforce to roll out marketing, e-commerce Einstein Copilots

Dig Deeper on Marketing and sales

Content Management
Unified Communications
Data Management
Enterprise AI