Smartsheet has thrown open the doors of its Model Context Protocol (MCP) server to three of the most powerful AI assistants on the market. As of June 11, 2026, enterprise customers can connect Microsoft Copilot, ChatGPT, and Google Cloud Gemini Enterprise directly to their live work management data—no static exports, no copy-paste roulette. The expansion builds on Smartsheet’s earlier integration with Anthropic’s Claude and introduces Smart Assist, an AI governance layer that promises to tame the chaos of enterprise AI adoption.
For Windows enthusiasts, the Copilot hook is the headline act. It means that Microsoft’s deeply integrated AI companion—already baked into Windows 11, Microsoft 365, and Edge—can now reach into Smartsheet to pull real-time project plans, resource allocations, and workflow statuses on demand. But the announcement is much bigger than one operating system. It’s a deliberate move to position Smartsheet as the neutral data backbone for the coming wave of multi-agent AI workflows, where employees juggle several assistants across different platforms.
What exactly is Model Context Protocol?
Smartsheet first unveiled MCP in late 2025 as a secure bridge between large language models and the company’s work management platform. Instead of forcing users to download CSV files or manually copy data into chat windows, MCP lets authorized AI models query Smartsheet’s API directly—with granular, admin-controlled permissions. Think of it as a bouncer at the data nightclub: only the right AI guests get in, and they only see the tables that management has approved.
Under the hood, MCP exposes a subset of Smartsheet’s REST API over a protocol that LLMs can consume naturally. When an AI assistant needs to answer a question like “What’s the status of the Q3 marketing campaign?” it sends a structured request to the MCP server. The server checks the AI’s identity and permissions, translates the query into an internal Smartsheet search or report, and returns the result in a format the AI can parse. The entire round-trip happens in seconds, and none of the raw data is stored by the AI provider unless explicitly allowed by the organization’s data retention policies.
The new Smart Assist component adds an essential layer of governance. Admins can define which sheets, reports, and dashboards each AI provider is allowed to access, set rate limits to prevent runaway API costs, and monitor all AI interactions through a dedicated audit log. Smart Assist also includes pre-built templates for common Copilot, ChatGPT, and Gemini configurations, reducing the time from “we should test this” to “let’s go live” from days to minutes.
Why the multi-AI expansion matters
Smartsheet’s initial MCP launch supported only Claude, reflecting its partnership with Anthropic. That exclusivity made sense as a proof of concept, but enterprise AI usage quickly fragmented. The 2026 Enterprise AI Adoption Survey from Gartner found that 73% of large companies now use at least two different AI assistants regularly, with Copilot and ChatGPT dominating the Windows-centric desktop while Gemini holds strong in Google Workspace shops. By adding all three giants simultaneously, Smartsheet eliminates the awkward “well, we only support Claude” conversation that IT teams dread.
For organizations that have standardized on Microsoft 365, the Copilot integration is transformative. An employee can ask Microsoft 365 Copilot in Excel or Teams, “Show me all tasks assigned to me this week from the Smartsheet project plan,” and get a live answer without ever leaving their workflow. Windows 11’s native Copilot sidebar can now surface Smartsheet items alongside local files and web results, blurring the lines between project management and the operating system itself.
ChatGPT Enterprise users gain the same real-time context within OpenAI’s secure business environment. A marketing manager could ask ChatGPT to draft a status update for a campaign based on the latest Smartsheet data, and the reply will reflect actual phase completion percentages, not yesterday’s snapshot. Google Cloud Gemini Enterprise customers, meanwhile, can connect Gemini’s reasoning to Smartsheet’s project tracking, enabling natural-language queries like “Which projects are over budget?” with direct sourcing from live financial fields.
Enterprise governance finally catches up to AI ambition
The Smart Assist layer is arguably the more important half of this announcement. As companies have raced to adopt AI assistants, security teams have been sounding alarms about data sprawl, shadow AI usage, and the risk of confidential information leaking into external models. Smartsheet’s approach confronts these fears head-on.
Smart Assist requires explicit authorization for every AI connection. An Azure AD or Okta group can be linked to a specific AI provider, and within that context, row-level and column-level security policies apply. For example, an HR manager using Copilot could see employee names and leave balances from a Smartsheet, while a contractor using the same Copilot instance might see only project task assignments without personal data. This isn’t just API key security—it’s full integration with existing identity providers and data classification frameworks.
Audit trails in Smart Assist log every query, including the AI model used, the user who initiated the interaction, the exact data returned, and a timestamp. Compliance teams can feed these logs into SIEM tools like Microsoft Sentinel or Splunk, turning AI interactions into auditable, reportable events. For heavily regulated industries like finance and healthcare, this traceability is non-negotiable.
Rate limiting and cost control are baked in at the admin level. Since MCP interacts with Smartsheet’s API, each AI query consumes platform resources. Smart Assist lets administrators set daily or monthly query caps per AI provider and per user, preventing both accidental cost explosions and denial-of-service scenarios. Early adopters in the Smartsheet Beta program reported that a single enthusiastic employee with a connected ChatGPT could generate tens of thousands of API calls in a day if left unchecked—Smart Assist puts guardrails around that enthusiasm.
Real-world scenarios: Copilot in Windows, ChatGPT, and Gemini at work
How does this play out on a typical Tuesday morning? Picture a project manager who starts her day by opening Copilot in Windows 11. Without launching a browser, she types: “Summarize today’s overdue tasks from Smartsheet.” Copilot, via MCP, reaches into the company’s Smartsheet instance, identifies three overdue items across two projects, and presents them in a concise list. She clicks one, and Copilot opens the corresponding Smartsheet row in her default browser. Total time: six seconds.
Meanwhile, a data analyst using ChatGPT Enterprise wants to understand a new dashboard. He uploads a screenshot of the Smartsheet dashboard as context (or simply references it by name) and asks, “Explain the variance in the Q3 marketing budget line.” ChatGPT, equipped with MCP access, retrieves the underlying sheet data and generates an analysis that references actual transaction records, not just the visual summary. The analyst can then ask follow-up questions like “Show me the individual expenses that caused the overrun,” and ChatGPT drills deeper into the source sheets.
In a Google-centric environment, a sales director using Gemini Enterprise asks, “Which deals in the pipeline have been stuck in the negotiation stage for more than two weeks?” Gemini queries the Smartsheet MCP server, which returns records from the sales tracking sheet filtered by stage and duration. The director gets a ranked list with dollar amounts and owner names, and with one click, she can jump into the relevant Smartsheet row to update notes or trigger a workflow.
These workflows aren’t hypothetical. Smartsheet’s announcement includes references to pilot customers who connected all three AI providers during a restricted beta. A large construction firm used Copilot to pull safety inspection data from Smartsheet into Microsoft Teams channels, slashing the time to close corrective actions by 40%. A global retailer connected ChatGPT to inventory tracking sheets, enabling store managers to ask natural-language questions about stock levels without learning structured query syntax.
The competitive landscape and Smartsheet’s strategic bet
Smartsheet’s MCP expansion arrives as the workplace platform market grapples with the AI agent revolution. Microsoft Loop, Notion, Coda, and Monday.com all offer varying degrees of AI integration, but most rely on either embedding a single assistant (like Notion AI) or using plugins that require manual data selection. Smartsheet’s protocol-based approach is more ambitious: it aims to make the underlying work data agnostic to the AI consumer.
Anthropic’s announcement of the open-source Model Context Protocol standard in late 2024 provided the technical foundation, but Smartsheet has pushed the concept into production at scale. By supporting multiple AI models through a single governance interface, Smartsheet is betting that enterprises will demand flexibility. The company doesn’t want to own the AI layer; it wants to own the truth layer—the single source of real-time work data that every AI assistant taps.
This positioning has implications for Microsoft too. Windows Copilot’s value grows exponentially when it can access more enterprise systems natively. A Copilot that only understands Office documents is useful; one that can reach into Smartsheet, Salesforce, and ServiceNow becomes a genuine business accelerator. Smartsheet’s MCP effectively does the heavy lifting of making Copilot smarter without Microsoft having to build bespoke integrations for every platform. That’s a win for joint customers, and it arguably strengthens Copilot’s ecosystem.
Challenges and the road ahead
No technology this new arrives without friction. The biggest concern is AI accuracy when dealing with complex, unstructured Smartsheet data. Smartsheet sheets can contain everything from rigid project plans to free-form notes, attachments, and formulas. Asking an AI to “find the deliverable that’s causing the delay” requires sophisticated filtering and reasoning, and early testers report that Copilot and ChatGPT occasionally misattribute tasks or misinterpret custom field logic. Smartsheet is working on prompt guidance and schema mapping tools to improve alignment, but the learning curve is real.
Data sensitivity remains a sticking point. Even with Smart Assist’s controls, some compliance teams will balk at granting any external AI provider live access to internal sheets. Smartsheet emphasizes that data never passes through its own servers—the AI providers connect directly to the MCP endpoint—but for the most secretive organizations, air-gapped environments and on-premises AI models may be the only acceptable path. Smartsheet acknowledges this and hints that an on-premises MCP gateway is under consideration for regulated customers.
Pricing has not been detailed, but industry watchers expect MCP Smart Assist to be a premium add-on, possibly bundled with Smartsheet’s Advance tier or sold as a separate subscription per connected AI provider. Small and medium businesses may find the cost prohibitive compared to simpler bot integrations. Smartsheet will need to balance enterprise-grade features with accessibility to avoid fragmenting its user base.
What this means for Windows users and the future of work
For the Windows community, the most immediate takeaway is that Copilot’s reach just got a lot longer. The assistant that lives in your taskbar is no longer a novelty that summarizes web pages and tweaks settings. It’s now a legitimate gateway into your company’s operational backbone, provided your IT team has connected the dots. The integration also demonstrates the strength of open protocols like MCP—a rare moment of cooperation in a fiercely competitive AI market.
Looking ahead, Smartsheet’s strategy suggests where the industry is heading. AI won’t be a single product you buy; it will be a layer that sits on top of your data, with multiple interfaces vying for your attention. The winners will be the platforms that make their data most accessible to the widest range of AIs, while still locking down security. Smartsheet’s MCP with Smart Assist is one of the first production-ready implementations of that vision.
The challenge for IT leaders is to turn this capability into a safe, productive advantage. That means investing in policies, training, and monitoring—not just flipping a switch. The employees who figure out how to wield a multi-AI command center with live data will outperform those stuck in manual mode. Smartsheet just handed them the keys. Now it’s up to the enterprise to drive.