M365 Adoption Lead | 2X Microsoft MVP |Copilot | SharePoint Online | Microsoft Teams |Microsoft 365| at CloudEdge
Excel’s Interactive canvas lets Copilot generate an interactive HTML dashboard from workbook data, so users can explore information through visualizations and controls instead of building every element by hand. It is relevant to Excel users who want a dashboard and are comfortable refining a generated result through natural-language instructions. Ami Diamond [MVP] demonstrates the workflow using Copilot in Excel on the web.
Interactive canvas creates a dashboard-style view from a description of the desired result. In the demonstration, the workbook contains usage data such as users, teams, countries, conversations, messages, and credits, and the requested output includes visualizations, key performance indicators (KPIs), and filters.
The output is HTML, rather than a conventional worksheet dashboard assembled by manually placing Excel charts and controls. That distinction matters when making changes: the demonstrated dashboard is not directly editable as a normal Excel layout. Instead, the user describes a change in Copilot and asks it to regenerate or adjust the dashboard.
The first result does not necessarily match every detail of the request. Here, the initial filters were not the dropdown controls the user wanted. A follow-up prompt changed them to dropdowns with multi-select, and another prompt requested a pie chart for the token breakdown. The workflow is therefore iterative: generate a first version, test it, then specify adjustments.
This approach is aimed at people who already have relevant data in an Excel workbook and want to explore or present it as an interactive dashboard. The example uses a usage-data workbook, but the basic task is to ask Copilot for a dashboard based on the data at hand.
The demonstrated starting point is Excel for the web with Copilot available. Before creating the dashboard, the publisher checks Copilot’s plug-in management area for a Visualization skill named Interactive canvas. The presence of that skill is the indicator used in the demonstration that the dashboard workflow is available in that environment.
The feature’s availability is still in rollout, so it may not appear for every Microsoft 365 customer. The demonstration takes place in the web version of Excel; the publisher notes that desktop availability may come in the future, but does not demonstrate it. No particular license or additional setup is identified in the demonstrated steps.
Start by opening the workbook in Excel for the web and checking that Copilot offers the Interactive canvas skill. The example checks this before writing the dashboard prompt. Use the workbook that contains the data you want to visualize, then give Copilot a concrete description of the output and its interactions.
In the example, the follow-up prompt replaces the filters with dropdowns that support multi-select. The dropdowns also provide options to clear the selection or select all. A separate prompt asks Copilot to replace the token breakdown visualization with a pie chart.
Testing the interactions is part of the process, not just a final design check. The publisher notes that a filter affected a list in an earlier attempt but did not do so in this example. That difference is a practical reason to click through the generated controls and confirm that the displayed information behaves as expected.
The clearest limitation in this workflow is the editing model. The generated HTML dashboard cannot be adjusted manually in the demonstrated experience; changes have to be requested through Copilot prompts. That means a user who wants a different chart, filter style, or layout should describe the change rather than expect to select and edit the dashboard element directly.
Also inspect the generated result instead of assuming that every requested component appears or behaves exactly as intended. The first dashboard did not show the desired filter style, and the example’s behavior differed from a previous attempt for one list. The user corrected the filter controls with a follow-up instruction, then made another targeted request for a different chart type.
For users who need direct, cell-level control over a dashboard, a conventional Excel approach using worksheet elements may be a better fit. Interactive canvas offers a prompt-driven alternative, but the demonstration does not establish that it replaces existing Excel dashboard methods. Treat the generated view as something to review and refine before relying on it.
In Excel for the web, open Copilot, check Manage plug-ins for the Visualization category’s Interactive canvas skill, and prompt Copilot to create a dashboard from your workbook data. Review it and use follow-up prompts to adjust the controls or charts.
In the demonstrated experience, the HTML dashboard cannot be edited manually. Request changes through Copilot prompts instead.
Yes. The example uses a follow-up prompt to request multi-select dropdown filters and another to change a token breakdown chart to a pie chart.
The demonstrated workflow uses Excel for the web. Desktop availability is mentioned as a possibility for the future, but is not demonstrated.