News
Lumo can draw the chart. You still need to define the number.
Proton’s August visualization feature is a useful reason to make the counting rules explicit before asking for a graph.
Announcement date: . Our context and analysis: Sep 6, 2026.
The useful part
Keep these three things in mind.
- Define the unit and calculation before requesting a visual.
- Separate chart quality from the decision to share data.
- Keep the checked source table beside the final graphic.
How this piece was prepared. Reporting from the linked primary announcement and current documentation, with our workflow analysis. No hands-on performance test.
What Proton announced
Proton announced on August 3, 2026 that Lumo can create charts, graphs and other visuals directly in chat. The announcement describes visual responses to supplied datasets and follow-up requests that refine the output. It also says Lumo may choose to produce a visual when that would help explain a response.
This expands the possible output from a conversation. It does not establish that a generated chart has used the right unit, calculation or comparison. We have reviewed the announcement and related documentation, not benchmarked the feature or audited a generated workbook.
Sources: Proton
Define one row before one chart
Before uploading a file, describe what a row represents: an order, a customer, a support message or a monthly total. Then name the measure and the time window. This is our recommended preparation, not a product requirement.
For an illustrative customer report, a count of messages is not automatically a count of customers. Someone who contacted support three times can contribute three rows. Ask for the proposed counting rule in plain language before asking for the final graphic, especially when the file contains duplicates, missing values or several currencies.
Treat data handling as a separate decision
Proton accompanies its product with a published Lumo security-model explanation. That document is a separate part of evaluating whether the service is appropriate for a particular dataset. A visualization feature announcement alone cannot answer every question about permitted data use, retention or organizational policy.
For an initial exercise, use a small non-sensitive dataset whose totals you can calculate independently. That isolates the presentation question from the decision to share confidential information. If the tool fits the task, review the applicable data controls before broadening the inputs.
Sources: Proton
Keep the working behind the picture
Ask for the aggregation used, the excluded rows and the table behind the visual. These are suggested review requests; confirm what the product actually returns. Reconcile the total with the original file, and inspect labels, units and any truncated axis.
A chart can be an excellent working draft without being ready for a board deck. Keep the checked data and the final image together so the next person can understand the number. The feature is most useful when it shortens the route to an inspectable explanation, rather than making an unexplained result look finished.
Sources & method
Reporting from the linked primary announcement and current documentation, with our workflow analysis. No hands-on performance test.
- Lumo can now turn your data into visuals ↗Proton
- Lumo security model: How Proton makes AI private ↗Proton
Sources checked Sep 6, 2026. Product capabilities can change; verify the current documentation before making a commitment.
Published by Pixel & Shelf. Prepared with AI assistance, with claims checked against the linked sources.
Our editorial approach Suggest a correction ↗