Salesforce Pioneers Generative AI Innovations for Developers with Einstein 1 Studio

Salesforce is advancing its generative AI capabilities with the launch of Einstein 1 Studio, coinciding with the TrailblazerDX developer conference.

Einstein 1 Studio equips developers with tools to customize the Salesforce Einstein Copilot generative AI assistant, which was recently released in beta. Although initially announced in 2023, its widespread availability begins today.

While Einstein Copilot focuses on user interactions, Einstein 1 Studio emphasizes the development of personalized user experiences. It features three primary tools: Copilot Builder, Prompt Builder, and Model Builder. Copilot Builder allows users to design custom AI actions for specific business tasks. Prompt Builder facilitates the creation and implementation of tailored prompts in workflows, while Model Builder offers the flexibility to build or import various AI models to cater to unique requirements.

“We want developers to have the tools to innovate with AI, ensuring they are equipped for an AI-first future, regardless of their coding expertise,” stated Alice Steinglass, EVP and GM of the Salesforce platform, during a TrailblazerDX press briefing.

Industry Pressures Surrounding Generative AI

At the TrailblazerDX event, Salesforce unveiled new research highlighting a disconnect between AI expectations and IT capabilities. According to the findings, 87% of IT professionals believe generative AI has met or exceeded its anticipated impact. This perception has heightened demands from management on developers to expedite AI implementation. However, 88% of IT professionals reported challenges in fulfilling all AI-related requests.

Steinglass emphasized that AI is not only transforming how developers create applications but also the nature of the applications themselves.

“Previously, developers had to hard-code every action, button, and workflow. AI is unlocking new possibilities, enabling dynamic conversations with customers and facilitating hyper-personalized experiences,” she explained.

Customization: The Core of Einstein 1 Studio

Organizations can now tailor the default Einstein Copilot using the Copilot Builder within Einstein 1 Studio.

“Copilot Builder allows companies to create custom AI-driven actions, configuring Einstein Copilot to suit their specific needs,” Steinglass noted. The actions generated through Copilot Builder leverage Salesforce’s developer tools, such as Apex, Flow, Heroku, and MuleSoft APIs, familiar to Salesforce developers. This integration empowers teams to enhance Einstein Copilot functionality across Salesforce and external systems.

An essential component of any AI copilot tool is the prompt, which is where the Prompt Builder tool comes into play. This low-code tool streamlines the process of crafting and refining AI prompts.

“Every interaction with a large language model (LLM) is driven by a prompt that delivers both data and instructions,” Steinglass clarified. “Prompt Builder lets customers experiment with prompts, testing and refining them while integrating directly with their trusted CRM and data cloud.”

Additionally, Prompt Builder enables developers to infuse AI into any Salesforce workflow, creating prompts for text fields, guiding processes, summarizing meetings, and embedding generative AI wherever necessary.

Flexibility with Model Builder

A central feature of Einstein 1 Studio is the Model Builder.

“Unlike other platforms that restrict businesses to a single LLM, our approach is open,” Steinglass highlighted. “The Model Builder allows connections to a diverse range of AI models.”

While organizations are not required to provide their models, they may choose to customize models for their specific employee or customer needs. Model Builder supports various AI platforms, including Anthropic, Cohere, Databricks, Google Cloud’s Vertex AI, and OpenAI.

“This flexibility is poised to empower a new wave of innovators in AI development,” she added.