
As a data engineer, the biggest win is how quickly Sundial takes me from a business question to something my stakeholders can actually use.
On the modeling side, I can describe a change to our semantic layer in plain language (a new metric, a dimension, a model tweak) and review the generated definitions before anything is applied. It takes most of the repetitive work out of maintaining the layer without giving up control over the final logic, so our metrics stay governed and reusable.
Where I've gotten the most mileage is Data Apps. I've used them to replace fixed vendor reports and stand up interactive dashboards for different teams: time-range selectors, breakdown dropdowns to re-slice by segment, and a "View SQL" button on every tile so anyone can see exactly how a number is computed. That transparency has made it far easier for non-technical stakeholders to trust the numbers, and I can build and iterate on a dashboard in a fraction of the time it used to take. Review collected by and hosted on G2.com.
Mostly small things. For complex or ambiguously defined questions I still like to double-check the generated query against the warehouse before I rely on it. The "View SQL" transparency makes that quick, and I'd love even tighter built-in validation so I can lean on the output with less manual checking. Data Apps are also newer and still maturing (deeper drill-downs and filtering are being polished), though never in a blocking way, and the team ships fixes fast. Review collected by and hosted on G2.com.