Smartsheet will integrate its Model Context Protocol (MCP) Server with ChatGPT, Microsoft Copilot, and Google Cloud Gemini Enterprise in June 2026, extending the collaborative work management platform’s AI assistant roster beyond its existing Claude support. The expansion comes with new governance controls designed to keep enterprise data secure and auditable as teams tap generative AI directly from their spreadsheets, projects, and automated workflows.

The Bellevue, Washington-based company revealed the update in a product roadmap announcement, positioning the MCP Server as a bridge between granular work management and the conversational AI interfaces millions of knowledge workers already use daily. By supporting four major AI ecosystems simultaneously, Smartsheet aims to eliminate the friction of switching between tools while maintaining centralized oversight over how sensitive business information is processed.

What the Smartsheet MCP Server actually does

Smartsheet introduced its MCP Server in early 2025 as an open connector that lets external AI applications query, update, and reason over Smartsheet data using the standardized Model Context Protocol. This protocol, which Anthropic originally developed to give Claude secure access to external tools, has rapidly become a cross-industry standard for linking large language models with business platforms.

Unlike conventional API integrations that require each AI vendor to build a custom connector, MCP creates a universal language. Smartsheet administrators configure which sheets, reports, and dashboards are exposed through the server, set row-level and column-level permissions, and apply role-based access. When an AI assistant like ChatGPT or Copilot receives a natural language request — “Show me the budget variance for the Q3 marketing campaign” — the assistant translates that into an MCP call, retrieves the precise data from Smartsheet in a governed fashion, and returns the result within the same chat interface.

This architecture means the AI never stores the underlying data permanently. Queries are executed in real time against live Smartsheet objects, and every interaction is logged for audit purposes. The June 2026 update extends this capability from Claude to ChatGPT, Copilot, and Gemini Enterprise, so organizations can pick the AI assistant their teams prefer without sacrificing data governance.

Why Smartsheet is betting big on a multi-AI strategy

Until now, Smartsheet’s AI narrative centered heavily on Anthropic’s Claude. The company was an early adopter of MCP and touted Claude’s 200,000-token context window as ideal for parsing complex project plans. But enterprise customers rarely standardize on a single AI provider. IT departments often field requests from finance teams that live in Microsoft 365 and want Copilot, while product teams rely on Google’s Vertex AI ecosystem and Gemini, and others prefer ChatGPT’s multimodal features.

By adding Copilot and Gemini alongside ChatGPT, Smartsheet acknowledges this fragmented reality. A marketing director can ask Copilot in Teams to summarize last month’s campaign spend without leaving their daily workflow. A supply chain analyst using Gemini in Google Sheets can pull real-time inventory levels from a Smartsheet inventory tracker. A legal team drafting contracts in ChatGPT can query compliance status from a Smartsheet-controlled log. Each scenario inherits the same Smartsheet permission model, so a user who only has view access to a sheet cannot request the AI to modify cells or download attachments.

This multi-assistant approach also future-proofs the platform. As Microsoft, Google, and OpenAI race to upgrade their models, Smartsheet can absorb those improvements without rebuilding integrations. The MCP layer insulates business logic from the underlying AI, so a Copilot update that introduces better natural language reasoning automatically benefits Smartsheet queries without a separate release.

Governance is the headline feature

The most critical piece of the June update is the enhanced governance framework. Smartsheet has long offered enterprise-grade controls — SOC 2 Type II, ISO 27001, HIPAA compliance — but extending them to AI interactions required a new layer of policy enforcement. The updated MCP Server introduces three specific safeguards:

  • Query approval workflows. Administrators can require that any AI-generated data modification — such as updating a status column or adding a comment — goes through a manual approval. The approval request appears in Smartsheet’s native workflow engine, where designated stakeholders can approve or reject the change before it commits.
  • Context-aware masking. When an AI assistant requests data, Smartsheet can automatically mask personally identifiable information, financial figures, or intellectual property based on classification tags applied to columns. The AI receives placeholder tokens instead of real values, and the unmasking happens only in the final response visible to the authorized user.
  • Audit trails with AI attribution. Every MCP interaction logs which AI assistant initiated the call, what query was made, which rows or cells were accessed, and how the data was transformed in the response. These logs integrate with SIEM platforms like Splunk and Microsoft Sentinel, giving security teams a complete forensics trail.

These controls address the primary enterprise objection to AI: loss of data sovereignty. Companies can now confidently let employees use any supported AI while maintaining a single source of truth inside Smartsheet.

Practical use cases across departments

Combine governed access with conversational AI, and the use cases multiply quickly. Here are several that Smartsheet highlighted in its roadmap presentation:

  • Project management triage. A portfolio manager types into Copilot: “Which projects are at risk this quarter?” Copilot reads the “Status” and “Health” columns from the PMO dashboard, cross-references them with due dates, and returns a prioritized list — without the manager opening Smartsheet. If the manager then asks, “Move the deadline for the Adams project by two weeks and notify the owner,” Copilot triggers the change via MCP, but only after the governance workflow gets approval from the portfolio lead.
  • Budget reconciliation in ChatGPT. A finance user pastes an invoice number into ChatGPT and asks, “Does this invoice exceed the approved budget?” ChatGPT uses MCP to locate the invoice’s associated project line in Smartsheet, compare the amount to the budget column, and reply. Because the finance user only has read access to budgets, the AI cannot accidentally create a new expense line.
  • Field-service coordination with Gemini. A service technician using a Gemini-powered mobile app queries, “What are the open tickets within 10 miles of my current location?” Gemini accesses the Smartsheet service log, filters by geography and status, and returns the address, priority, and customer notes. The technician never sees tickets from other regions due to Smartsheet’s row-level security.

These scenarios illustrate how AI becomes a seamless front end for structured work data, while Smartsheet remains the system of record that enforces business rules.

Competitive landscape and what this means for Windows and Microsoft 365 shops

Smartsheet’s expansion with Copilot is particularly notable for the 400,000-plus organizations using Microsoft 365. Microsoft has been aggressively embedding Copilot across its ecosystem — in Excel, Teams, SharePoint, and Power Platform — often touting its own Loop components as the bridge between structured and unstructured work. Smartsheet’s MCP integration essentially offers an alternative: instead of migrating project data into Loop or Excel for Copilot to access, teams can keep it in Smartsheet and make it available through MCP.

This positions Smartsheet as a Copilot-enabled data hub that competes directly with Microsoft Lists and Planner integration but with deeper project management and portfolio-level capabilities. For IT administrators managing hybrid environments, the benefit is clear: they can let business units choose Smartsheet for its dynamic sheets, Gantt charts, and automation while still empowering those users with Copilot’s conversational interface from anywhere.

Google Workspace customers get a similar advantage with Gemini Enterprise. Rather than copying Smartsheet data into Sheets for AI analysis, Gemini can reach directly into governed Smartsheet objects. This keeps the source data authoritative and avoids versioning nightmares.

Community reaction and early feedback

While the full product rollout is months away, initial responses from Smartsheet’s customer advisory board and social media channels point to cautious optimism. Power users who have already built extensive automations with Smartsheet Bridge and Data Shuttle are eager to see whether the MCP layer reduces the complexity of those workflows. Instead of building a complex data shuttle to export snapshots into a warehouse for AI consumption, they can surface the same data on demand through an assistant.

Some skeptics, however, worry about the learning curve. Configuring the MCP Server requires mapping Smartsheet security roles to AI-specific permissions, and misconfiguration could expose sensitive columns. Smartsheet is addressing this with a policy simulator that lets admins test AI access scenarios before enabling production traffic. Additionally, the approval workflows add a process step that may slow down urgent requests — a trade-off that compliance-heavy industries like financial services will accept more readily than fast-moving tech startups.

What comes next: the roadmap beyond June 2026

Smartsheet has hinted at further enhancements planned for late 2026 and early 2027, including:

  • Agentic workflows. Allowing AI assistants not just to read and write cells but to orchestrate multi-step automation — for example, taking meeting minutes in Teams via Copilot, extracting action items, creating Smartsheet tasks, assigning owners, and pinging them in Slack.
  • Graph-based data connections. Expanding MCP to traverse relationships between sheets, dashboards, and external data sources, so an AI can answer a question like, “Show me the resource capacity for every project under the North America VP” without pre-building a report.
  • Bring-your-own-LLM support. For organizations with sovereign cloud requirements or custom fine-tuned models, Smartsheet plans to let customers connect their own AI endpoints through the MCP framework, routing queries to their private instances of models like Llama or Azure OpenAI Service.

The company is also working on a natural-language policy builder where admins can describe governance rules in plain English — “Marketing can see campaign budgets but not salaries” — and have the system automatically generate the corresponding MCP permissions.

Preparing for the June rollout: three action items for Windows enterprise admins

For IT teams managing Windows-based environments that rely on Smartsheet, the next few months offer a window to prepare:

  1. Audit your Smartsheet permission model. The MCP Server will faithfully replicate every workspace, folder, and sheet-level permission. Inconsistent or overly permissive access controls will propagate directly to AI assistants, so clean up role assignments now.
  2. Engage your data classification project. The context-aware masking feature depends on column-level tags. Start labeling columns that contain PII, financials, or IP so that the masking rules can be applied immediately upon activation.
  3. Run a pilot with a single AI assistant. If your organization already uses Copilot heavily, begin a limited pilot with a small team of power users once the feature becomes available in the Smartsheet beta channel. Measure both user satisfaction and any increase in Smartsheet adoption before scaling to ChatGPT or Gemini.

Smartsheet has indicated that documentation, SDK updates, and a series of live configuration workshops will roll out in the weeks before the June 2026 general availability date. The company is also updating its trust center with detailed MCP-specific security architecture diagrams to help compliance teams evaluate the setup.

The bigger picture: governing AI access, not blocking it

The June 2026 MCP Server update reflects a broader enterprise trend away from blanket AI bans and toward governed enablement. For years, companies siloed their data in platforms like Smartsheet while employees copied and pasted snippets into AI chatbots, bypassing security entirely. By opening a governed channel through MCP, Smartsheet acknowledges that AI usage is inevitable and provides a safer path.

The availability of four major AI assistants under one governance framework may become a model for other SaaS platforms. ServiceNow, Salesforce, and Atlassian are all exploring MCP-based integrations, and Smartsheet’s multi-vendor approach could accelerate the standard. For Windows-centric enterprises, the ability to keep data in Smartsheet while interacting through Copilot in the familiar Microsoft interface blurs the line between standalone SaaS and native productivity tools — a convergence that promises less context-switching and more time spent on actual work.

As organizations continue to grapple with AI’s productivity promise versus its compliance peril, Smartsheet’s updated MCP Server stakes a clear position: the path forward is secure, auditable, and choice-driven.