Salesforce unveils data & AI tools to tackle governance gaps

Salesforce has announced an expansion of its platform, introducing a set of new data governance, security, and semantic capabilities designed to support enterprises scaling artificial intelligence technologies across their operations.

According to research from RAND, over 80% of artificial intelligence projects in organisations do not succeed in delivering value, with poor data quality, insufficient governance, and integration challenges among the chief causes. Salesforce's new features aim to address these obstacles, providing what the company describes as the "scaffolding for the Agentic Enterprise"-a model in which humans and AI agents collaborate across workflows, decisions, and interactions.

Features and capabilities

The Salesforce Platform now includes three central capabilities in its approach to trusted AI: context and accuracy, built-in trust and compliance, and open, unified systems. These capabilities are intended to ensure that AI outputs are grounded in unified business data and that security and compliance are integrated into every workflow, while avoiding vendor lock-in through open interoperability.

A key component is a new enterprise-wide metadata layer. This aims to provide structure to both complex and unstructured data, so that AI outputs can be more consistent and explainable for business users.

Data Cloud Context Indexing

Salesforce has introduced the Data Cloud Context Indexing, a new indexing pipeline that enables AI agents to interpret unstructured content, such as contracts, diagrams, and tables, through a business-focused perspective. By surfacing specific details from large and fragmented datasets, the platform allows users to receive accurate information more efficiently. One use case provided by Salesforce is that a field engineer could upload a schematic and receive step-by-step guidance from an agent to troubleshoot technical issues, reducing the required time for resolution.

Data Cloud Clean Rooms

The company is now offering Data Cloud Clean Rooms as a generally available product. This allows organisations to securely share, collaborate on, and analyse data without the need to move or expose raw data, using zero copy connectivity. This model is intended to reduce compliance and security risks, as well as storage costs. For example, global banks can use Clean Rooms to compare transaction patterns and detect fraud without revealing sensitive customer information to one another. The service is also integrated with AWS Clean Rooms to enable privacy-enhanced collaboration with partners.

Tableau Semantics

Another major announcement is Tableau Semantics, an AI-driven semantic layer integrated with Data Cloud. It interprets raw data into business language, aiming to unify data and metadata across different business domains and platforms. Tableau offers a Customer 360 Semantic Data Model (SDM) as a standard package, which simplifies the process of modelling and ensures consistency in metrics and business definitions. Salesforce has stated that it is working with industry partners such as Databricks, dbt Labs, and Snowflake to promote standardised semantics across analytics and AI, helping organisations avoid inconsistencies in key business definitions, such as revenue calculations.

MuleSoft Agent Fabric

As companies accelerate AI adoption, they are experiencing new challenges associated with an increasing number of AI agents-a phenomenon described as "agent sprawl." Salesforce is responding with MuleSoft Agent Fabric, an orchestration platform allowing businesses to register, manage, and govern AI agents developed both internally and externally. This is designed to prevent fragmentation and enable seamless collaboration between different AI-powered functions, such as inventory management, pricing updates, and fraud detection, even across different business units or systems.

Embedded Security and Informatica Acquisition

Salesforce has also embedded new AI-powered security and compliance features across its platform, including integrations with industry security providers such as CrowdStrike and Okta. These integrations are intended to provide proactive threat identification and compliance management within organisations' AI environments.

A planned acquisition of Informatica is expected to bring additional metadata intelligence capabilities to Salesforce's data catalogue, integration, governance, privacy, and Master Data Management portfolio, subject to the closing of the transaction.

Salesforce perspective

"At Salesforce, we believe AI is the most powerful enabler of business transformation today. But for it to deliver on its promise, it must be grounded in deep business context to power the true force multipliers: intelligent automation and actionable insights," said Rahul Auradkar, EVP & GM, Unified Data Services. "That's why we're building a unified foundation that harmonises all of a company's structured and unstructured data to surface these insights and fuel this automation, all grounded in the governance and security needed to operate with confidence. With these innovations, we're helping our customers move faster, scale responsibly, and unlock the full potential of the Agentic Enterprise."

Customer views

AAA Washington has been implementing Salesforce to reinforce its sales and service operations. Jim Ryan, Chief Information Officer of AAA Washington, commented, "AAA Washington has recently implemented Salesforce to our support sales and service organisations. With a unified data foundation powered by Data Cloud and MuleSoft, AAA Washington is pursuing a roadmap that unlocks the promise of a 360° view of members. Instead of struggling with siloed systems, we will ultimately deliver faster, more accurate, and more relevant answers when it matters most. The partnership with Salesforce is helping to transform roadside, insurance, and travel member experiences and helps to fulfil the AAA Washington mantra of 'helpful tech, human touch.'"

In the healthcare sector, Andrew Chang, Chief Marketing Officer at UChicago Medicine, emphasised the importance of trustworthy AI: "AI in healthcare must be built on trust. By establishing a unified foundation for data, we can ensure each patient interaction is reliable, accurate, and meaningful, while enabling staff to dedicate more time to the complex, human-centred care that defines our commitment to patients."