Smartsheet will allow enterprise customers to connect their own instances of Microsoft Copilot, ChatGPT, and Google Cloud Gemini Enterprise into governed workflows, starting June 11, 2026. The Bellevue, Washington-based company announced that its Model Context Protocol (MCP) server will serve as a secure bridge between external AI assistants and Smartsheet’s work management platform, enabling AI-powered automation while retaining strict access controls and audit trails.

The move marks a significant expansion of Smartsheet’s AI capabilities, giving enterprises the freedom to choose preferred large language models (LLMs) without sacrificing the governance frameworks that regulated industries demand. Instead of relying solely on Smartsheet’s own AI features, organizations can now bring their own AI keys and let employees use familiar tools like Copilot or ChatGPT directly within Smartsheet sheets, reports, and dashboards—all mediated by the MCP server.

Model Context Protocol: The Open Standard Driving the Integration

At the heart of this announcement lies the Model Context Protocol, an open standard initially developed by Anthropic to connect AI models with external data sources and tools. MCP provides a universal way for applications to expose their functionalities to LLMs, much like a USB-C connector for AI integrations. Instead of building custom integrations for every AI platform, software vendors can implement an MCP server once and instantly become compatible with any MCP-compliant client.

Smartsheet’s implementation of an MCP server is a strategic play to position its work management platform as an AI-agnostic hub. By adhering to the protocol, Smartsheet avoids vendor lock-in and allows enterprises to swap AI providers as their needs evolve. As of mid-2026, MCP has gained considerable momentum, with major players like OpenAI, Google, and Microsoft supporting or integrating with the protocol in various capacities.

The Smartsheet MCP server acts as a gatekeeper. When an employee prompts Copilot or ChatGPT to, say, summarize project status from a Smartsheet sheet, the AI assistant sends a request to the MCP server. The server then checks permissions, retrieves only the data the user is allowed to see, and returns a structured response. That response fuels the AI’s natural-language output without exposing raw data beyond the governance envelope.

Bringing Your Own AI: Copilot, ChatGPT, and Gemini Enterprise

The June 11 announcement specifically names three AI services: Microsoft Copilot, OpenAI’s ChatGPT, and Google Cloud Gemini Enterprise. Each comes with distinct strengths, and Smartsheet’s MCP integration lets enterprises mix and match.

Microsoft Copilot integration is particularly noteworthy for the Windows-centric workforce. Copilot is embedded in Windows 11, Microsoft 365, Edge, and Teams, and many knowledge workers already use it daily. By connecting Copilot to Smartsheet’s MCP server, users can query their work data without leaving the Copilot pane in Windows or a Microsoft 365 app. A project manager might ask Copilot, “What are the top three overdue tasks in the Q3 launch sheet?” and get an answer pulled directly from Smartsheet, respecting row-level security and data residency policies.

ChatGPT integration brings OpenAI’s widely adopted conversational AI into the mix. Enterprises that have built internal workflows around GPT models can now point those instances at Smartsheet data. This is particularly useful for teams that already use ChatGPT for drafting reports, analyzing trends, or generating status updates. With the MCP server, ChatGPT can access real-time data instead of relying on stale exports or manual copy-paste.

Google Cloud Gemini Enterprise rounds out the offering for organizations standardized on Google Workspace and Google Cloud. Gemini’s multimodal capabilities and deep integration with BigQuery and Looker may appeal to data-heavy teams that use Smartsheet alongside Google’s analytics stack.

Crucially, Smartsheet is not providing these AI models itself. Enterprises must bring their own licenses and API keys. The MCP server handles authentication and authorization, ensuring that only approved AI instances can communicate with Smartsheet and that all interactions are logged.

Governance at the Core: Security, Auditing, and Data Residency

For enterprises in finance, healthcare, government, and other regulated sectors, AI integration often stalls on governance concerns. Smartsheet’s MCP server addresses these head-on with a comprehensive governance layer.

Role-Based Access Control (RBAC): The server inherits Smartsheet’s existing user and group permissions. If a worker running an AI query lacks access to a particular sheet, column, or row, the server automatically filters that data out of the AI’s response. This prevents accidental data leaks through AI summarization or question-answering.

Comprehensive Audit Logs: Every AI-driven request—who made it, which AI service was used, what data was accessed, and what response was returned—gets logged in an immutable audit trail. Security teams can review these logs to spot unusual patterns or ensure compliance with internal policies.

Data Residency and Sovereignty: Smartsheet’s MCP server operates within the customer’s chosen data region. AI prompts and responses are processed there, and data never traverses to unexpected geographies. This is essential for GDPR, HIPAA, and other regulatory frameworks.

Administrative Controls: IT admins can set organization-wide policies via Smartsheet’s admin center. They can disable specific AI services, restrict usage to certain user groups, or enforce approval workflows before AI access is granted. These levers give enterprises the confidence to roll out AI assistance gradually.

How It Works in Practice: A Governed Workflow

Consider a typical scenario: A marketing operations manager needs to update a campaign budget sheet. She opens the Copilot sidebar in Microsoft Edge and types, “Analyze the ‘Q4 Campaign Tracker’ sheet in Smartsheet and flag any line items that exceed the allocated budget by more than 10%.”

Copilot sends a tool-use request to Smartsheet’s MCP server. The server authenticates the user via her Microsoft Entra ID, verifies she has view access to the tracker sheet, retrieves only the relevant columns (campaign name, allocated budget, actual spend), calculates the variance, and returns a structured JSON payload. Copilot then renders the answer in natural language: “Three line items exceed budget by over 10%: Digital Ads (+15%), Print Collateral (+22%), and Event Sponsorship (+18%).”

The entire interaction takes seconds and is logged with a timestamp, user identity, AI service used, and data fields accessed. The manager can act on the insight immediately without switching apps or worrying about data sprawl.

For teams using ChatGPT, the flow is similar, just triggered through a ChatGPT interface or API. Enterprises can build custom GPTs that hook into Smartsheet’s MCP server, allowing specialized assistants for sales forecasting, resource allocation, or risk tracking. Google Gemini users can do the same within Google Workspace or Cloud Console.

The Windows Angle: Copilot as the Front Door

For Windows users, the Copilot integration may prove the most adopted path. Windows 11 ships with Copilot integration at the OS level. Users can summon Copilot with Win+C and issue natural-language commands. With Smartsheet’s MCP server connected, those commands can extend directly into work management data.

IT administrators configuring Windows devices for enterprise can deploy the Smartsheet MCP connector as part of the Microsoft 365 admin center or Intune. Once deployed, users can authenticate once and then use Copilot across all their Microsoft endpoints—Windows, Mac, iOS, Android—to interact with Smartsheet data. This aligns with Microsoft’s broader vision of Copilot as a unified AI assistant across the Microsoft ecosystem.

Smartsheet has long offered deep integration with Microsoft 365, including real-time co-authoring with Excel, embedding in Teams, and syncing with Outlook calendar. The MCP-based Copilot connection is a natural next step, marrying AI assistance with established collaboration workflows.

Industry Response and Competitive Landscape

Smartsheet’s announcement lands amid a flurry of MCP adoptions across the enterprise software world. Work management competitors like Asana, Monday.com, and Wrike have all introduced AI features, but Smartsheet’s bring-your-own-model (BYOM) approach stands out. Instead of forcing customers into a proprietary AI assistant, Smartsheet lets them leverage their existing investments in Copilot, ChatGPT, or Gemini.

This strategy may appeal to large enterprises that have already negotiated enterprise agreements with Microsoft, OpenAI, or Google. It also insulates Smartsheet from the rapid pace of LLM development. If a customer wants to swap from ChatGPT to a new open-source model in six months, they can do so without changing their Smartsheet workflows—as long as the new model supports MCP.

The MCP ecosystem itself is expanding. As of 2026, dozens of software vendors—from database companies to CRM platforms—have launched MCP servers. This network effect benefits Smartsheet, because enterprises can compose complex automations that span multiple tools. For example, a sales team might use Copilot to query Smartsheet for deal status, then tell the same Copilot to pull related emails from Outlook and generate a customer summary—all within a single governed workflow.

Implementation Timeline and Availability

Smartsheet plans to roll out the MCP server in public preview on June 11, 2026, with general availability expected later in the year. The feature will be available to all Enterprise and Premier plan customers at no additional cost; however, customers must supply their own AI service licenses. Initial release supports Microsoft Copilot, ChatGPT, and Google Cloud Gemini Enterprise, with plans to expand to other MCP-compatible models based on customer demand.

Administrators will configure the MCP server through Smartsheet’s admin console, where they can upload API keys, set access policies, and review audit logs. Setup guides and documentation will be available on Smartsheet’s help portal on launch day. Smartsheet will also offer professional services engagements for complex deployments.

What This Means for the Future of Work Management

Smartsheet’s MCP play signals a maturation of enterprise AI. Where early AI integrations often involved simple chatbots or single-purpose features, the MCP server model envisions AI as an ambient layer across all work tools. With governed access to work management data, AI assistants become less of a novelty and more of a productivity multiplier.

The emphasis on governance cannot be overstated. As AI adoption accelerates, regulators worldwide are sharpening their focus on data privacy and algorithmic accountability. Smartsheet’s decision to build auditability and access control into the MCP server from day one positions it well for compliance-conscious customers.

For Windows enthusiasts and IT pros, the Copilot integration is a window into a future where the OS-level AI assistant becomes the primary interface for all enterprise software. Rather than navigating through Smartsheet’s web app, users can stay in their flow and let Copilot fetch and manipulate data on their behalf. This aligns with Microsoft’s “Copilot everywhere” narrative and deepens Smartsheet’s value within the Microsoft ecosystem.

Conclusion

Smartsheet’s June 2026 addition of Microsoft Copilot, ChatGPT, and Google Cloud Gemini Enterprise through its Model Context Protocol server represents a thoughtful convergence of flexibility and control. Enterprises can now deploy the AI models their teams prefer, while IT retains rigorous governance over data access, residency, and usage. The integration underscores MCP’s growing role as the lingua franca for AI-tool connectivity and positions Smartsheet as an agnostic hub for the AI-powered enterprise. Windows users stand to gain a particularly seamless experience, with Copilot serving as a natural front door into governed work management workflows. As the preview rolls out, all eyes will be on adoption metrics and real-world feedback from heavily regulated industries—the ultimate test of whether governed AI can deliver on its promise without compromising security.