Enabling autonomous AI agents at scale

Salesforce’s plan to acquire Informatica will unleash a powerful trifecta of technologies, making it easier for organizations to benefit from a new, human/digital hybrid workforce

Deploying autonomous AI agents at scale is poised to transform business operations. Enterprises across all industrial sectors are eager to leverage these agents – working alongside humans – to boost productivity, efficiency, and customer experience. However, to unlock the full value of the digital labor opportunity, it’s imperative that companies empower AI agents with broad access to organizational tools and data – and do so without sacrificing security or incurring massive integration costs. The recently announced plan by Salesforce to acquire Informatica is good news for enterprises as they address this significant challenge.

Overcoming deployment barriers

In Rise of agentic AI: How trust is the key to human-AI collaboration, the Capgemini Research Institute projects AI agents could generate up to $450 billion in economic value by 2028, through revenue uplift and cost savings across the surveyed countries.

But from their interviews with 1,500 executives, Capgemini researchers discovered only 14 percent of organizations have moved beyond pilot projects to partial or full-scale deployment of these agents. Trust is a key barrier, as those surveyed cited ethical concerns, lack of transparency, and a limited understanding of agentic AI capabilities. But organizational readiness – including the creation of an effective governance system – is also hampering secure, scalable deployments.

Salesforce is one of the leading technology companies helping enterprises to deploy autonomous agents, and it is taking steps to help organizations overcome these barriers. In the spring of 2025, Salesforce announced a major play to strengthen its capabilities in the form of an $8 billion deal to acquire Informatica. As a longtime Salesforce partner, Capgemini believes this is an important development that will enable Salesforce to deliver AI agents that can operate with intelligence, context, and confidence across the modern enterprise.

Key assets, working together, will enable this.

Agentforce. The Salesforce approach starts with Agentforce – the company’s flagship AI agent platform. Agentforce integrates natively with an organization’s existing applications, data, and business logic so agents can securely take action across the enterprise – handling complex tasks automatically while working in tandem with human teams.

Early deployments of Agentforce have already demonstrated substantial gains. For example, companies using Agentforce have cut customer service case handling time by double-digit percentages and allowed AI agents to autonomously resolve the majority of simple support requests. At scale, these AI agents handle high-volume, repetitive tasks such as answering FAQs, processing form submissions, or triaging support tickets. This frees up human agents to focus on higher-value work.

Salesforce recently enhanced this solution with the Agentforce Command Center, which enables business leaders to monitor and control their AI agents’ activities in real time. This level of visibility and governance addresses critical hurdles to scaling AI agents across the enterprise.

Anthropic’s Model Context Protocol. To enable its AI agents to access diverse systems, tools, and data across the client’s organization, Salesforce has embraced Model Context Protocol (MCP) – an open integration standard from Anthropic. This addresses a major pain point in the AI deployment process – namely, that custom integrations, each using custom code and requiring unique maintenance processes, do not scale.

MCP eliminates the need for developers to build a custom integration every time agents need to connect to external systems, APIs, databases, and services. The result is faster development, lower integration costs, and the freedom to mix-and-match AI models with a wide variety of tools and data sources. MCP’s model-agnostic open standard – often referred to as “the USB-C of AI” – means businesses avoid vendor lock-in and encourages a broad ecosystem of integration. Salesforce’s decision to adopt MCP enables Agentforce agents to seamlessly interface with a vast and growing universe of enterprise systems and cloud services – without requiring custom code, and without compromising on security.

MCP-native agents. When Salesforce released Agentforce version 3 in mid-2025, it introduced built-in MCP interoperability. What’s more, more than 30 launch partners provide MCP integrations – spanning cloud platforms (AWS, Google Cloud), content and collaboration tools (Box, Notion), payments (PayPal, Stripe), data and AI services (IBM, Writer), and more. This means Agentforce can accomplish a vast variety of tasks.

The Salesforce vision is clear: to enable an open ecosystem in which Agentforce-powered AI agents can plug-and-play into business applications and services, regardless of source and with minimal setup. This represents a major leap forward in what these agents can do autonomously.

The Informatica toolset. The effectiveness of AI agents – no matter how intelligent or well integrated – is only as good as the data on which they operate. With its plan to purchase Informatica, Salesforce will acquire important enterprise-grade tools for data integration, data quality and cleansing, master data management, granular data governance and privacy controls, and real-time data observability across complex hybrid and multi-cloud environments.

From a business perspective, this acquisition will inject a powerful dose of data integrity, context, and governance into Salesforce’s AI ecosystem, ensuring Agentforce agents have access to clear, trusted, and actionable data. Enterprises will be able to track where data comes from, how it’s transformed, and how it’s used. Organizations will avoid mistakes due to using outdated or inconsistent data. And companies will deploy AI agents, confident that they will not run afoul of regulatory requirements or privacy laws.

A powerful trifecta

Agentforce, MCP, and Informatica form the three pillars of an AI-driven enterprise: an agent platform to act, a protocol to connect, and a data ecosystem that informs. Organizations that leverage all three will be well positioned to achieve unprecedented levels of automation and insight – transforming their enterprise into a smarter, more agile business in which humans and agents can collaborate seamlessly to enrich customer experiences and drive growth.

For many enterprises, this will make the vision of autonomous agents a practical reality. AI agents, working fluidly across systems, will handle routine processes in customer service, sales, marketing, IT, and finance. This digital workforce will answer questions, generate reports, update records, and flag issues – autonomously, and in real time. This will free up humans to focus on strategic, creative, and relationship-oriented work – activities at which humans excel – while supervising AI as needed.

Capgemini is excited by this trifecta and looks forward to working with its Salesforce clients to enable the ongoing value opportunity agentic AI represents. As a Salesforce partner for 17 years and one of the company’s global top five strategic partners, Capgemini offers its clients expert knowledge of the Salesforce platform, the experience of more than 3,000 AI specialists and 50,000 AI-enabled engineers, strong integration capabilities, and sector-specific expertise in multiple industries. Assets include the Capgemini Agentforce Factory – a hub for clients to explore real-world applications through interactive demos, hands-on training, and expertise guidance.

For more information, please contact: andy.forbes@capgemini.com