Salesforce has expanded AgentExchange into a single marketplace that brings together AppExchange, Slack Marketplace, and the wider Agentforce ecosystem, as it pushes to make enterprise AI deployments easier to discover, activate, and govern across customer-facing and internal workflows.
Unveiled at TDX 2026, the new AgentExchange combines 10,000 Salesforce apps, more than 2,600 Slack apps, and more than 1,000 agents, tools, and Model Context Protocol servers in one catalogue. Salesforce said the marketplace is designed to let customers discover, buy, activate, and manage solutions across Salesforce and Slack from a unified cataloguing layer. The marketplace is live at AgentExchange.
The marketplace update did not arrive on its own. Salesforce paired it with a wider set of platform changes under its new Headless 360 architecture and an expansion of Agent Fabric, the company’s control layer for discovering, governing, orchestrating, and observing AI agents across an enterprise environment. Agent Fabric now includes broader discovery for external platforms including Amazon Bedrock, Microsoft Foundry, and GoDaddy, while new governance and orchestration controls are intended to help companies manage how agents behave in production.
Taken together, the announcements suggest Salesforce is moving beyond the initial launch phase of Agentforce and into the harder work of enterprise deployment. The company is no longer only selling the idea of agents embedded in CRM workflows. It is now trying to solve the operational issues that emerge once organisations begin managing multiple agents, models, marketplaces, and surfaces at once.
“You are not pulling your customers into your mobile app or website anymore. You’re pushing your experience and brand and meeting them wherever they are. That is the power of Headless 360 and the experience layer,” said Janani Narayanan, VP, Unified Profile & Knowledge at Salesforce.
That line captures an important part of the company’s direction. Salesforce is treating channels, interfaces, and marketplaces less as separate products than as distribution points for the same underlying logic, data, and AI workflows. In that model, the question is not only which agent a company builds, but where that agent can be safely deployed, what business context it can access, and how consistently it behaves across surfaces such as Slack, voice, web, and mobile.
The change also reflects a wider industry problem. Enterprises have spent the past year assembling AI layers across service, sales, and operations, often using a mix of internal tools, large-model providers, workflow platforms, and specialist vendors. That has created a fragmented landscape in which discovery, governance, and cost control are becoming as important as model performance. Salesforce’s answer is to make AgentExchange the commercial front door while Agent Fabric becomes the control plane behind it.
There is financial weight behind that strategy. In its most recent quarterly results, Salesforce said Agentforce annual recurring revenue had reached $800 million, up 169% year on year, with 29,000 deals closed and nearly 20 trillion tokens converted into more than 2.4 billion agentic work units. Those figures help explain why the company is now tightening the commercial and operational plumbing around the platform.
Rivals across the CX and enterprise software market are moving in the same direction. Recent launches from Genesys and NiCE have also focused on agentic AI that can resolve work rather than merely assist human users. What sets Salesforce’s latest move apart is the emphasis on consolidation. By unifying marketplaces and adding tighter governance over a multi-vendor environment, it is trying to make agent deployment look less like a collection of pilots and more like an enterprise software estate.
The immediate task for customers will be practical rather than conceptual: deciding which agents belong in which workflows, what data and permissions they should inherit, and how much variation they can tolerate once those systems are live. Salesforce is betting that a larger marketplace will be more useful if it is paired with stronger orchestration, tighter controls, and a clearer route from discovery to production.
You must be logged in to post a comment.