Salesforce Unveils New Plan to Boost AI Agent Trust

Salesforce has launched a significant expansion of its platform capabilities, focusing on data, governance, security, and semantics to strengthen trust in AI agents. This move addresses a critical industry challenge, as widespread adoption of agentic AI depends on establishing reliability and security. Many AI initiatives currently fall short due to issues like poor data quality and insufficient governance frameworks.

Building confidence in AI agents is essential for their mainstream use. Salesforce’s Agentforce product represents a major commitment to this technology, yet surveys indicate that trust remains a barrier. For instance, one study found that over 80% of AI projects fail to deliver expected value, often because of data inconsistencies and integration problems. Another recent analysis identified compliance, data quality, and governance as the primary obstacles to scaling AI agents in marketing.

To help businesses scale agentic AI reliably, Salesforce is rolling out several platform-wide enhancements:

Data Cloud Context Indexing introduces a new pipeline within Salesforce Data Cloud that enables AI agents to understand unstructured content, such as contracts, diagrams, and tables, from a business perspective. This capability pulls precise information from large, disconnected datasets, helping users obtain accurate answers more quickly.

Data Cloud Clean Rooms allow organizations to securely share, collaborate on, and analyze data without moving or exposing raw information. By using zero-copy connectivity, these clean rooms avoid duplicating sensitive datasets, which lowers security risks, eases compliance burdens, and reduces storage expenses. Companies can also work safely with external partners through Salesforce’s integrated connection with AWS Clean Rooms.

Tableau Semantics is an AI-driven semantic layer built into Salesforce Data Cloud. It converts raw data into business-friendly terms, speeding up time to value. Tableau now offers a pre-built Customer 360 Semantic Data Model that brings together data and metadata from various sources, simplifies modeling, ensures consistent metric governance, and supplies the business context required for dependable AI and business intelligence insights. Supporting this effort, Tableau is helping develop the open semantic interchange standard.

MuleSoft Agent Fabric addresses the challenge of “agent sprawl,” where disconnected workflows, redundant automations, and compliance gaps arise as AI agents multiply across teams and systems. This solution provides a centralized hub for registering, orchestrating, and governing every AI agent, no matter where it was developed.

Embedded AI security and compliance features are being integrated across the Salesforce Platform, making it easier to maintain security and meet compliance requirements. New integrations with CrowdStrike and Okta will enhance threat detection and compliance management.

Additionally, Salesforce’s planned acquisition of Informatica will incorporate Informatica’s data catalog, integration, governance, quality, privacy, and Master Data Management tools into the Salesforce Platform. This will create a unified data architecture for agentic AI, supplying the metadata intelligence enterprises need to ensure their AI agents operate safely, responsibly, and at scale.

The potential for agentic AI across various industries is immense, but achieving it is a gradual process. Salesforce had already introduced trust and governance measures earlier, and this latest announcement is likely just one step in an ongoing series of developments aimed at fostering dependable AI ecosystems.

(Source: MarTech)