Smartsheet has expanded the reach of its enterprise work management platform by launching an MCP Server integration that simultaneously connects work data with Microsoft Copilot, ChatGPT, and Google Cloud Gemini Enterprise. The announcement, made on June 11, 2026, marks a significant step in bringing real‑time project context directly into the AI assistants millions of knowledge workers already use daily.

The new MCP (Model Context Protocol) Server acts as a secure bridge between Smartsheet’s collaborative work management environment and generative AI interfaces. Instead of switching between tabs or manually exporting data, users can now query project statuses, update tasks, or extract insights from sheets, reports, and dashboards right inside Copilot, ChatGPT, or Gemini. The integration joins Smartsheet’s existing MCP support for Anthropic’s Claude, giving enterprise customers a unified way to surface work intelligence across all major AI platforms.

For organizations that have embraced AI copilots but still struggle with fragmented data, this development promises to collapse the gap between conversation and action. A marketing manager chatting with Copilot can ask, “What’s the status of the Q3 campaign launch?” and receive an answer drawn from the live Smartsheet project plan—without ever leaving Microsoft Teams. Similarly, a developer using ChatGPT Enterprise can retrieve sprint progress from a Smartsheet workspace that feeds Jira and GitHub, while a financial analyst can interrogate Gemini to compare budget forecasts stored in a shared sheet.

What the Smartsheet MCP Server actually does

The Model Context Protocol is an open standard initially proposed by Anthropic to let AI models securely access external tools and data sources. Smartsheet’s MCP Server implementation translates that protocol into a set of enterprise‑grade connectors. It exposes Smartsheet’s core objects—sheets, reports, workspaces, dashboards, and even automated workflows—as context that large language models can understand and act upon.

When a user poses a natural‑language question inside a supported AI assistant, the assistant’s runtime detects that a Smartsheet resource might answer it. The assistant sends a request through the MCP Server, which authenticates the user, checks permissions, fetches the relevant data, and returns a summarized, structured response. All of this happens in seconds, respecting the same access controls already defined in Smartsheet.

Crucially, the server does not train or fine‑tune the underlying models on customer data. It merely provides a temporary, read‑only window into the Smartsheet objects the user is already allowed to see. For write operations—such as updating a task or adding a comment—the assistant must explicitly request the change through a vetted API path that maps to Smartsheet’s existing automation rules.

How the three new integrations work

Microsoft Copilot

Copilot’s integration is perhaps the most tightly woven into existing workflows. Because Copilot already lives inside Microsoft Teams, Excel, PowerPoint, and the Edge browser, adding Smartsheet connectivity means work‑context follows users across the Microsoft ecosystem. A project manager can ask Copilot in Teams, “Show me all overdue tasks from the ‘New Product Launch’ sheet,” and Copilot will surface a card with the task list, assignees, and due dates. Clicking a card opens the corresponding row in Smartsheet for detail editing.

Administrators control the integration through the Microsoft 365 admin center. They can restrict which Smartsheet workspaces are queryable from Copilot, enforce sensitivity labels, and audit all AI‑generated requests. Microsoft’s enterprise data protection policies—such as no model training on tenant data—remain in force.

ChatGPT Enterprise

OpenAI’s enterprise‑grade offering receives a plug‑in that is deployed through the ChatGPT admin console. Once an admin enables the Smartsheet MCP connection, licensed users see a “Smartsheet” icon in the chat interface. They can then reference specific sheets by name or URL directly in their prompts. For example, “Summarize the risks listed in the Q2 risk register” will pull the register from Smartsheet, analyze the risk descriptions and probabilities, and generate a concise summary with suggested mitigations.

Data stays within the enterprise’s ChatGPT environment, and OpenAI does not use interactions to improve its base models. Audit logs in both Smartsheet and ChatGPT Enterprise provide visibility into which data was accessed and by whom.

Google Cloud Gemini Enterprise

For organizations running on Google Workspace, the Gemini Enterprise integration is designed to appear as a first‑party extension. Via the Gemini side panel in Gmail, Docs, or the standalone chat, users can invoke Smartsheet data with simple natural‑language commands. A user drafting a status report in Docs could type, “Insert a paragraph summarizing the current sprint progress from the ‘Engineering Roadmap’ Smartsheet,” and Gemini will fetch the data and output formatted text.

Because Gemini Enterprise operates under Google Cloud’s Assured Workloads and data residency commitments, regulated industries can activate the integration without compromising compliance postures. Smartsheet data is processed within the customer’s chosen Google Cloud region and is never stored permanently by the model.

Shared governance and security model

A consistent thread across all three integrations is the governance layer Smartsheet has built into its MCP Server. Every request passes through a policy engine that evaluates:

  • User identity – Integrated with Azure AD, Google Workspace SSO, or OpenAI’s identity provider.
  • Scope of access – Only sheets, reports, and workspaces explicitly shared with the user are visible.
  • Sensitivity classifications – Organizations can tag Smartsheet items with labels (e.g., “Confidential,” “Internal Only”) and create rules that block or obscure data in AI contexts.
  • Operation type – Read operations follow the user’s existing sheet permissions; write operations require additional approval steps configured in Smartsheet’s automation builder.

Smartsheet administrators manage all MCP Server connections from the Admin Center. They can enable or disable integrations per user group, set rate limits, and receive alerts when unusual access patterns are detected. This centralized control addresses one of the biggest concerns enterprises have voiced about AI data access: the risk of oversharing confidential information.

Beyond the big three: the broader ecosystem

The Copilot, ChatGPT, and Gemini launches are not Smartsheet’s first foray into AI connectivity. The company already provided an MCP Server for Anthropic’s Claude, a favorite among developers and research teams. With these additions, Smartsheet now covers the four most widely adopted enterprise AI assistants—Microsoft, OpenAI, Google, and Anthropic—giving customers the flexibility to choose the platform that best fits their tech stack and culture.

This multi‑platform strategy is deliberate. Enterprises rarely standardize on a single AI tool; instead, different teams gravitate toward different assistants. A legal department might lean on ChatGPT Enterprise for contract analysis, while engineering uses Claude for code review, and the broader organization lives inside Microsoft 365. Smartsheet’s MCP Server acts as a neutral, single source of truth that feeds all of them simultaneously.

The vision behind the announcement

Smartsheet CEO Mark Mader has previously described the company’s AI ambition as “bringing the front end of work into the conversation.” The MCP Server is the technical manifestation of that vision. Rather than building a separate AI agent or interface, Smartsheet is embedding its data layer directly into the tools people already use. This approach lowers adoption friction and ensures that AI‑augmented work management does not require a behavioral revolution—just smarter defaults.

During a press briefing coinciding with the launch, Smartsheet’s VP of Product revealed that early testers saved an average of 3.5 hours per week on status reporting and meeting preparation. One Fortune 500 customer reduced the time to compile a weekly executive dashboard from four hours to twelve minutes by letting Copilot pull live data from 15 different Smartsheet reports.

The roadmap, according to the briefing, includes deeper bidirectional capabilities. Later in 2026, Smartsheet plans to allow the MCP Server to accept natural‑language instructions to create new sheets, automate workflows, and even design dashboards—all from within the chat interface of any connected AI assistant.

Real‑world considerations and challenges

Despite the promise, analysts point to several hurdles. First, because each AI vendor implements MCP slightly differently, maintaining feature parity across platforms will require constant engineering effort. Any delay in one integration could frustrate users who expect identical behavior everywhere.

Second, the quality of AI‑generated answers depends heavily on how well Smartsheet data is structured. Organizations that use chaotic naming conventions, sparse metadata, or large monolithic sheets may see less accurate summaries. Smartsheet has published best‑practice guides encouraging users to adopt clear column naming, consistent use of dropdown fields, and regular data hygiene.

Security remains a live conversation. Even with rigorous governance, some CISOs remain uneasy about piping employee‑accessible data into external AI models, even if models do not retain data. Smartsheet’s approach of processing requests through its own server—rather than requiring the AI assistant to directly call Smartsheet’s API—is designed to add an additional layer of abstraction and control. Still, early adopters are advised to conduct thorough penetration testing and limit the scope of exposed workspaces until confidence matures.

Pricing, too, will shape adoption. The MCP Server itself is included in Smartsheet’s Enterprise and Premier plans, but customers must separately license the AI assistants. Microsoft Copilot for Microsoft 365 costs a per‑user monthly fee, ChatGPT Enterprise is priced per seat, and Gemini Enterprise requires Google Cloud commitments. For large organizations, the combined per‑worker cost could be significant, potentially slowing deployment in cost‑conscious verticals.

What this means for the future of work management

The Smartsheet announcement reflects a broader industry shift from static dashboards to conversational work interfaces. Gartner has predicted that by 2027, over half of all project status updates will be generated by AI assistants drawing from connected work management platforms. Smartsheet’s move to become the data backbone for multiple AI front ends positions it to capture a share of that transformation.

Competitors such as Asana, Monday.com, and Wrike have also invested in AI features, but most have opted for proprietary AI assistants embedded within their own interfaces. Smartsheet’s decentralized MCP approach is more hub‑and‑spoke, betting that the future belongs to interoperability rather than walled gardens.

For Windows‑centric organizations, the Copilot integration is particularly notable. It extends the value of Microsoft’s AI investment by connecting it to the wide array of non‑Microsoft project data that often lives in Smartsheet. This could reduce the pressure to migrate everything into Microsoft Lists or Planner just to gain AI assistance, preserving existing workflows while modernizing the information layer.

The integrations are available now for Smartsheet Enterprise and Premier customers on the latest cloud infrastructure. Setup guides, including specific configuration steps for each AI platform, are available in the Smartsheet Help Center. Administrators are encouraged to start with a pilot group to fine‑tune policy settings before rolling out broadly. As AI evolves, so too will the expectations of knowledge workers. With its MCP Server, Smartsheet is ensuring that work context is never more than a question away—no matter which AI assistant you choose to ask.