Brillaint comparison. The time-to-insight gradient (60min down to 5min) really underscores how the analyst's value proposition is shifting - less time cleaning data, more time framing the right questions. What stuck with me is that Tableau still won on polish for stakeholder presentations, which suggests AI tools haven't cracked the "last mile" of translating insights into persuasive narraitves. Ive seen analysts spend 70% of their time on formatting for exec dashboards, so freeing up that bandwidth for deeper analysis feels huge. Curious if the speed gains hold up when dealing with messy real-world schemas vs the relatively clean dataset used here.
They definitely still have some room to improve on more "real world" datasets that are messier and more complicated. As I teased in the comments above, we'll have an article coming next week using Google Antigravity to do an analysis with a Snowflake database.
Nice write up. This is a space I've been looking in to as well. The whole "vibe analytics" space is growing rapidly but it's still unclear on the actual utility of them.
I'm curious, did you evaluate other tools that didn't make the list? Thinking tools like Hex, Plotly Studio, Nao, etc.
I've also been getting in to Claude Code lately so I wonder how it would perform, especially with the new Opus 4.5 model.
I did evaluate Google Antigravity as well, but for connecting to Snowflake and performing analysis by actually writing queries. Teaser: it's coming in a write up next week.
I haven't used Claude too much. My Antigravity experiment I used the Gemini 3 Pro models and it performed pretty well.
I'm going to continue to look at more tools and evaluate potential use cases in analytics, so I appreciate you sharing those. They might get worked into a future writeup.
Yeah there's starting to be a lot of these types of tools popping up. I also tried Julius and thought it was really good. One of my biggest areas of focus is the "last mile analytics" which really brings everything together into the "so what". Julius has an Executive Summary feature that does a pretty good job for that and something that lot of these tools don't even offer.
Agreed, Julius is a pretty cool tool. It's definitely the most focused for data analysis, of all the ones I tested. It does a good job in that niche. I do need to test how it works with a live database, since for this experiment I just used a CSV
Brillaint comparison. The time-to-insight gradient (60min down to 5min) really underscores how the analyst's value proposition is shifting - less time cleaning data, more time framing the right questions. What stuck with me is that Tableau still won on polish for stakeholder presentations, which suggests AI tools haven't cracked the "last mile" of translating insights into persuasive narraitves. Ive seen analysts spend 70% of their time on formatting for exec dashboards, so freeing up that bandwidth for deeper analysis feels huge. Curious if the speed gains hold up when dealing with messy real-world schemas vs the relatively clean dataset used here.
Thank you!
They definitely still have some room to improve on more "real world" datasets that are messier and more complicated. As I teased in the comments above, we'll have an article coming next week using Google Antigravity to do an analysis with a Snowflake database.
Nice write up. This is a space I've been looking in to as well. The whole "vibe analytics" space is growing rapidly but it's still unclear on the actual utility of them.
I'm curious, did you evaluate other tools that didn't make the list? Thinking tools like Hex, Plotly Studio, Nao, etc.
I've also been getting in to Claude Code lately so I wonder how it would perform, especially with the new Opus 4.5 model.
Thanks!
I did evaluate Google Antigravity as well, but for connecting to Snowflake and performing analysis by actually writing queries. Teaser: it's coming in a write up next week.
I haven't used Claude too much. My Antigravity experiment I used the Gemini 3 Pro models and it performed pretty well.
I'm going to continue to look at more tools and evaluate potential use cases in analytics, so I appreciate you sharing those. They might get worked into a future writeup.
Yeah there's starting to be a lot of these types of tools popping up. I also tried Julius and thought it was really good. One of my biggest areas of focus is the "last mile analytics" which really brings everything together into the "so what". Julius has an Executive Summary feature that does a pretty good job for that and something that lot of these tools don't even offer.
Agreed, Julius is a pretty cool tool. It's definitely the most focused for data analysis, of all the ones I tested. It does a good job in that niche. I do need to test how it works with a live database, since for this experiment I just used a CSV