I took on a small weekend project to code a Databricks AI/BI Dashboard that shows my sales numbers. Given that I hadn’t coded in a while, I decided to give Genie Code a run. Below I reflect on my experience.
First of all, I was impressed by Genie’s ability to, not only understand coding, but also the practical aspects of sales metric: it intuitively knew to organize calls, call-connects, and meetings-booked into a sales funnel chart. With a small prompt, I had a decent dashboard in a matter of minutes.
Genie built the data sources and connected them to the dashboard’s widgets. For simple graphs, this was a painless and effortless process. After brewing my tea, I returned to a slick-looking dashboard with a sales funnel, two pie charts showing calls made per Account Executive (AE) and per account called. At the bottom, I had a bar-chart showing the number of calls I made each week on one axis and a line-chart showing the cumulative duration of all my calls (in seconds).
Agentic coding isn’t flawless yet: Genie got confused about the pie charts and some point, and started to over-engineer the input datasets to circumvent a trivial problem. This required my intervention. The downside(?) was that I had to think for myself again. The upside is that I now actually understand how AI/BI dashboards are built on a technological level.
I’ll definitely default to Genie Code for expanding my dashboard with additional metrics. I want to create a quarterly page that shows my performance across quarters and highlights my strengths as well as areas I need to work on.
Once I have something polished, I’ll share it with my team and hope the company adopts it. Given that Databricks is heavily leveraged at Databricks1, I hope my creations can empower our BDR-teams across the planet.
You don’t say… ↩︎