Exploring 4 Distinct Features of Salesforce’s Agentforce 2.0

“At Salesforce, our sole focus is on Agentforce.” This was the declaration made by Salesforce CEO Marc Benioff in September as he introduced the concept of AI agents. Recently, the company demonstrated the significance of this statement by launching an extensively enhanced version: Agentforce 2.0.

Within this update, four key features stand out prominently:

• A comprehensive library of pre-configured skills and workflow integrations designed for swift customization.
• The option to implement Agentforce directly within Slack.
• Upgrades to the Atlas Reasoning Engine.
• The introduction of the Agentforce Testing Center.

Let’s delve into these features and explore their practical applications.

Skills and Integrations Library

This library encompasses essential business tools such as CRM, Slack, and Tableau, along with integrations from AppExchange partners. These pre-developed skills simplify the creation of customized AI agents by removing the burdensome and intricate issues tied to integrations. The library includes pre-configured Slack Actions, such as Create Canvas and Message Channel, allowing teams to develop agents tailored for their specific Slack workflows.

Here are a few applications:

• The CRM Skills enable the creation of agents that can automate and enhance interactions with customers. They can assist with lead nurturing, sales coaching, scheduling service engagements, and more. For instance, the Sales Development skill can craft agents to nurture leads based on set criteria, while the Sales Coaching skill allows agents to join sales calls, providing real-time feedback to enhance representatives’ performance.
• The MuleSoft integrations allow organizations to leverage Agentforce across their entire operational landscape. For example, it can link Agentforce to ERP systems to streamline order fulfillment or connect with marketing automation platforms to initiate tailored email campaigns based on user behavior.
• AppExchange enhances Agentforce functionality by enabling users to access specialized skills developed by partners.

Agentforce within Slack

This feature positions Slack as the primary hub for employees to engage with Agentforce. Users can retrieve information, initiate workflows, and oversee tasks without the need to switch between various applications.

Agentforce can also automate operations within Slack. For example, Accenture utilizes it to provide staff with timely updates regarding significant accounts. This aids in automating document searches and offers suggestions for resolving business challenges. Additionally, it streamlines the proposal creation process and fosters collaboration by providing updates in real-time, enabling document retrieval, and generating proposals.

Explore further: SharkNinja begins its Salesforce AI adventure

Agentforce also harnesses Slack Enterprise Search, learning from discussions in Slack channels and direct messages. This allows agents to deliver more relevant insights and act based on the specific context of conversations.

Upgraded Atlas Reasoning Engine

The Atlas Reasoning Engine has been refined with sophisticated reasoning and retrieval augmented generation (RAG), empowering agents to address more intricate, multistep inquiries while providing highly accurate responses.

For instance, RBC employs Atlas in conjunction with Data Cloud and Agentforce to assist financial advisors. The Data Cloud consolidates all client data into a singular view, allowing advisors to access a complete understanding of their clients’ needs. Atlas can respond to complex questions from clients by applying multistep reasoning and leveraging RBC’s robust data and business logic, ensuring the accuracy of its answers.

As an old Russian proverb states, “Trust, but verify.” The enhanced RAG functions enable users to validate information by providing inline citations with agent responses. Citing specific sources utilized enhances trust in the AI’s outputs and assists users in verifying accuracy.

Agentforce Testing Center

The Testing Center offers businesses a platform to assess their AI agents prior to deployment. This guarantees that agents comply with established protocols, remain factual, and operate effectively. It helps mitigate risks and ensures consistent performance while furnishing tools that simplify testing, auditing, and scaling agents.

Implementing and Advocating

Salesforce utilizes this technology within its own operations, reporting that agents at help.salesforce.com now manage 83% of customer support requests independently, leading to a 50% reduction in human escalations following implementation just two weeks ago.

“As a CEO, I now find myself not only overseeing people but also managing agents,” remarked Benioff. “There exists a genuine agentic layer around the platform today. It is not merely a futuristic vision; it is happening at this moment.”