Smartsheet unveiled vsActual, a standalone AI-native planning and execution product, at its ENGAGE 2026 customer conference. The product turns strategy documents into a living plan and then tracks whether live work stays aligned with it. It is supported by a new Execution Graph built on 20 years of workflow intelligence, millions of projects and nine petabytes of data. For business leaders, the point is direct: faster task completion does not guarantee delivery of strategy.
How vsActual and the Execution Graph work
vsActual is designed to read across systems teams already use instead of forcing migration into one project tool. Smartsheet lists Oracle, SAP, Workday, Salesforce, Jira and ServiceNow among financial, HR, customer and service platforms it connects. The starting point is the strategy itself, taken from presentations, budget documents or spreadsheets and converted into goals, timelines and constraints on spending and staffing capacity. New requests are then assessed against those priorities before approval.
Scenario planning sits at the center of the mechanism. Leaders can review budget and capacity consequences of a decision before selecting a path, and the chosen scenario becomes the plan of record. Agents then compare ongoing activity with that approved plan, identify emerging delays or resource conflicts, and assemble executive summaries from underlying data. Agents cannot change the plan independently, and recommendations include reasoning for leaders to review. General availability and pricing were not specified, with customers directed to register for updates.
The Execution Graph supplies the business context behind those comparisons. It maps relationships from individual tasks through projects and programs to strategic goals, including dependencies, shared capacity, decision ownership, permissions and approvals. Smartsheet says this goes beyond collecting documents for a chatbot because operational relationships and governance requirements are part of the model. The release does not identify a newly trained foundation model or publish comparative benchmarks. Existing architecture combines its intelligence layer and knowledge graph with foundation models through Azure OpenAI and Amazon Bedrock.
What this means for enterprise AI adoption
For companies using AI, the practical change is a shift from isolated project assistance to plan-level control. A slipping milestone that looks manageable alone can become critical when several funded initiatives depend on the same team, and connecting schedules, capacity and priorities could expose that conflict earlier than separate dashboards. Small firms may use the living plan to test new requests against limited budgets and staff, while large organizations gain a shared reference across fragmented systems. Value still depends on completeness and freshness of connected records.
The surrounding platform release adds automation that reduces repeated setup and checking. Agentic Workflows turn natural-language instructions into recurring work, such as weekly project summaries or risk and dependency checks delivered without repeating the request. The Creation Agent builds sheets, dashboards, reports and automations from a plain-language goal, while the Data Quality Agent flags sheet problems by severity on demand or on schedule. Native integrations cover ServiceNow, Microsoft Dynamics 365, HubSpot and Workday, with an Integration Agent in early access for GitHub, Zendesk, Airtable and Monday. com.
The marker to watch is whether vsActual moves from registration to deployment with measurable alignment between plans and delivery. Relevant signals include named enterprise users, published availability dates, and evidence that agents flag resource conflicts earlier than current dashboards. Governance credentials such as ISO/IEC 42001 certification, the Australian IRAP assessment, Security Event Streaming to Splunk and Microsoft Sentinel, and Data Classification under Smartsheet Safeguard set the baseline. Correctness of recommendations will still need proof in live operation.
