Salesforce Enhances Agentforce with New AI Agent Command Center

In a world where artificial intelligence is rapidly transforming business operations, Salesforce stands out with a significant update to its Agentforce platform. With the introduction of a new command center dedicated to AI agents, the publisher is strengthening its offering to optimize the user experience and sales automation. Their latest developments aim to ensure effective supervision of the lifecycle of AI agents, while integrating innovative technologies to meet the growing needs of businesses. How will this advancement transform the CRM landscape and customer interaction?

On the occasion of Salesforce World Tour Paris, Olfa Kharrat, director of AI and Agentforce product management, presented these recent developments which promise to redefine industry standards. This article explores in depth the new features of the platform, as well as their implications on the market, adding insights on the integration of third-party agents.

What’s new in Agentforce: towards better observability

The update to Agentforce 3 is not just a technical development, it marks a strategic shift in the way Salesforce approaches the deployment and monitoring of AI agents. The new command center will allow companies to monitor the performance and behavior of AI agents in real time, a crucial aspect to ensure customer experience quality.

A dashboard for AI agent adoption

The newly launched dashboard will allow users to measure several key adoption metrics:

• Interaction rate with AI agents
• User feedback
• Operational costs
• Availability and latency of services

This feature comes with notable improvements to the testing center, which evaluates agents across thousands of use cases to ensure optimal performance. The goal is clear: to offer an iteration loop that ensures continuous improvement of AI agents, providing developers with recommendations based on observed performance.

Technical Specificities of Agentforce

Agentforce 3 introduces interesting technical features:

• Support for advanced models, such as Claude Sonnet and Gemini, to optimize response speed.
• A provision of an interface facilitating the assembly of tests and instructions via Agentforce Studio.
• Observability of agents’ reasoning processes, allowing complete transparency into their actions.

In short, these innovations aim to establish a closer connection between users and AI agents, while improving operational efficiency. Feature

Agentforce and Interoperability: A Pragmatic Vision One of the major focuses of the Agentforce update lies in its approach to interoperability, particularly with regard to agent communication protocols such as the Model Context Protocol (MCP) and the A2A protocol. Salesforce is committed to ensuring seamless communication between different AI agents while maintaining strict control over the data. One-way support to start

Currently, Salesforce’s MCP solution focuses on one-way integration. This means that only the MCP client can access external data and actions, raising questions about the company’s long-term strategy:

How can we foster greater collaboration between AI agents? Does this strategy risk limiting innovation? Respected analyst Rebecca Wettemann points out that the success of these technologies depends not only on the tools provided by vendors, but also on the ecosystem they manage to create. It is crucial that Salesforce work to make their tools easily accessible to developers to improve integration. Steps Toward Two-Way IntegrationTo refine this strategy, Salesforce is considering several steps:

Creating a registry for agents to register.

Developing a card for each agent to facilitate mutual recognition.

• Establishing secure communication between agents, thus enabling data exchange.
• This process, however, takes time. To date, only 10% of Salesforce customers have deployed AI agents in production, demonstrating that there is still a long way to go.

Execution

Description

Step 1

1. Create an Agent Registry
2. Step 2
3. Deploy Agent ID Cards

Step 3

Additionally, Salesforce must be proactive in communicating the benefits of AI for

sales automation

, highlighting case studies and testimonials from satisfied customers.

The challenges of integrating third-party agents

• Interoperability remains a key point for the future of Agentforce. As Salesforce considers communication between third-party agents, it will be vital to master the security and ethical aspects of this integration:
• Ensure the protection of user data in a multi-agent environment.
• Develop interoperability standards that are accepted by the publishing community.

Promote strong collaborations with other market players to enrich the Agentforce ecosystem.ChallengePotential solution

Limited adoption of AI agents

Training and awareness programs

1. Restricted interoperability
2. Establishing strategic partnerships
3. Security Concerns