Chloe Maraina is a powerhouse in the world of business intelligence, known for her unique ability to transform cold, complex big data into compelling visual narratives that drive corporate strategy. With a deep background in data science and a forward-looking vision for systems integration, she has become a leading voice on how enterprises can bridge the gap between siloed information and actionable AI. Today, she joins us to discuss the radical shift in the AI landscape brought about by the Model Context Protocol (MCP) and what it means for the future of the digital workplace.
Our discussion centers on the emergence of MCP as a universal standard that allows AI agents from different ecosystems—like those from Anthropic and OpenAI—to communicate seamlessly within a single interface. We explore the concept of “super agents” that handle the heavy lifting of orchestration, the elimination of the exhausting “swivel-chair” workflow that plagues modern sales teams, and the strategic evolution of Slack into a comprehensive operating system for the enterprise.
How do you see the Model Context Protocol changing the way data is shared across competing platforms like OpenAI and Anthropic?
The Model Context Protocol, or MCP, has effectively become the lingua franca of AI agents, serving as a vital handshake between disparate systems. By treating this open standard with the same level of importance as APIs during the microservices era, Salesforce is finally realizing a vision that co-founder Marc Benioff has championed since 2020, even before the massive $27.7 billion Slack acquisition was finalized. This protocol allows application data to be exposed safely to AI agents from outside the ecosystem, including those developed by Vercel, Anthropic, and OpenAI. It breaks down the walls of the “walled garden,” ensuring that a marketing agent can talk to a sales agent regardless of who built them. For the enterprise, this means interoperability is no longer a luxury but a baseline expectation for any serious AI deployment.
With the introduction of these new standards, Salesforce is positioning Slackbot as a “super agent.” What does that mean for the average employee trying to navigate their daily tasks?
The transition of Slackbot into a “super agent” is essentially an AI orchestration play designed to simplify the human experience by handling the background complexity of task-management. Instead of a person having to remember which specific agent handles CRM updates and which one pulls data from Tableau, Slackbot acts as the overarching intelligence that understands and tracks the functions of all other task-based agents. You can simply issue a plain-language order—like asking for a visualization of sales trends—and Slackbot figures out the routing and execution without you ever leaving the chat window. This effectively removes the cognitive load of managing a dozen different AI tools, letting the software do the heavy lifting of coordination. It is a significant step toward a future where we interact with our tools through natural conversation rather than complex menu navigation.
We often hear about the “swivel-chair” problem in enterprise environments. How do these new agentic integrations specifically address the friction found in sales and marketing workflows?
The “swivel-chairing” effect occurs when an employee has to constantly jump between Slack, Tableau, and Salesforce CRM just to complete a single workflow, which leads to massive productivity leaks and data latency. By embedding MCP integrations directly into the platforms people use most, these updates allow salespeople to trigger CRM actions and view complex data visualizations without ever switching tabs. This is particularly transformative for pipeline management; instead of having a forecast that is only accurate at one fixed point in time each week when the manager demands an update, the data flows in real-time. Automation aggregates the outcomes of sales meetings as they happen, ensuring that the CRM is a living, breathing reflection of the business. It turns the chore of data entry into a background process that happens naturally as part of the conversation.
There is often a fear that implementing sophisticated AI requires a massive engineering effort. How accessible is this new functionality for the average system administrator?
One of the most impressive claims regarding this latest set of Slackbot functionality is that it requires almost zero engineering lift. Salesforce has built the bridge between the MCP client and the MCP server on the backend, meaning an administrator doesn’t need a team of developers to get these tools running. If an admin is capable of turning their phone onto airplane mode, they have the technical skill required to manage these integrations via a simple slider-switch dashboard. This democratization of AI setup means that even small organizations can deploy agentic tools that previously would have required months of custom coding. It’s about moving the power of AI out of the lab and into the hands of the people who actually run the business operations.
If Slack becomes the primary interface for all of Salesforce’s functionality, what does the future look like for the traditional Salesforce UI that many have used for decades?
There is a real possibility that the traditional Salesforce interface we’ve known for years might eventually take a backseat to more intuitive, voice-activated, or chat-based channels. Slack CTO Parker Harris has suggested that if users find Slack and its various input channels so efficient that they never feel the need to return to the old UI, that is actually a positive outcome for the company. This shift represents a fundamental upgrade in how we work, moving away from clicking through endless tabs and toward a more fluid, integrated experience. The focus is no longer on the “home” of the data, but on the utility of the interface where the work actually gets done. While the core CRM database remains the heart of the operation, the way we interact with it is clearly evolving into something much more conversational and less rigid.
With so many new bots and platforms being released every week, many CIOs are feeling overwhelmed. How should leadership determine which AI tools to prioritize?
Right now, the level of confusion among enterprise tech buyers and CIOs is incredibly high because they are being bombarded with competing announcements and overlapping capabilities. It is crucial for vendors to be prescriptive, clearly defining when a customer should use Slackbot versus when they should build custom agents in Agentforce. Leaders need to look for use cases in their existing workflows that will actually deliver measurable efficiencies rather than just chasing the latest “shiny” AI tool. The goal should be to identify where the “yin and yang” of MCP and AI can solve specific bottlenecks, like real-time data aggregation or automated scheduling. Ultimately, the winners will be the organizations that focus on orchestration—using a platform that can manage a fleet of agents rather than trying to manage each one individually.
What is your forecast for the future of AI orchestration in the enterprise?
I believe we are entering an era where orchestration becomes the ultimate competitive battleground for every major tech vendor. Within the next few years, the raw power of generative AI models will be a commodity, and the real value will lie in how effectively a platform can coordinate those models to solve complex business problems. We will see the “super agent” model become the standard, where humans act as directors while a central AI orchestrator manages a diverse ecosystem of specialized bots. This will lead to a dramatic reduction in manual administrative work, potentially freeing up thirty percent or more of a knowledge worker’s day for higher-level strategic thinking. The companies that master this protocol-driven interoperability today will be the ones defining the workflow of the next decade.
