Make Narrative more polished and professional
Summary: Chris Wallingford shares feedback on the initial use of Narrative by their clients. While clients appreciate the tool for data validation during design, they find its text generation problematic in professional contexts. Issues such as incorrect casing, informal language, and misinterpretation of future data as anomalies deter them from sharing Narrative with end users. Consequently, Narrative remains a tool for designers and is not exposed to consumers to avoid negative perceptions from executive leaders.
Our clients have started using Narrative. Initial feedback is generally positive, but only to a point in the data pulling to presentation process. While pulling data together it's a nice to have for designers to see validation of what they're seeing. When they prepare content for a presentation, they might use the Narrative as a starting point for their own presentation notes (though copy/paste is a challenge with Narrative). What they don't feel comfortable doing is sharing dashboards to end users with Narrative enabled. Narrative often uses inappropriate casing and reads like a text message.
The designer my create values for "FY26 Revenue" and "FY27 Revenue" and Narrative will talk about how "fy26" is doing compared to "Fy27".
The designer might put calendar month names on the axis of a widget and Narrative will talk about the outliers of "july, august, and september" and "the pronounced December increase".
And the Narrative can't understand that some months in FY27 are in the future, and instead calls out "the absence of data" as being "a significant anomaly, possibly indicating a data gap or operational pause."
These kinds of responses reflect poorly on the dashboard designers and negatively impact Executive leaders' perceptions of their work. As a result, Narratives is "reserved for designer use" and not exposed to consumers.
