The full story
- Industry
- Retail & Consumer
- Use case
- BI & Analytics, AI Apps
- Warehouse
- Snowflake
- Published
- October 27, 2026
The company had spent four years and hundreds of hours building its reporting in a legacy BI tool. Double Wood moved to Sigma on Snowflake and rebuilt 12 of those dashboards with Sigma Migration Skills and Claude Code. The team queries a Sigma sales agent directly from the dashboard and can drill down to SKU level detail.
Weeks to hours – Time to convert one BI dashboard into Sigma
12 dashboards – Complex BI conversions in two weeks
80% – Of each conversion completed by the agent before review
01
The Challenge
Double Wood Supplements is a top 20 vitamins, minerals, and supplements brand and has grown sales exponentially in the last five years. Reporting changes within the existing infrastructure led to inefficiencies in scaling at the pace they needed to run at.
Double Wood’s BI tool held the company’s reporting, and it made the source of a number hard to check. When a figure looked wrong, answering the question meant opening Snowflake and reading the tables directly. Tracing a number that way took an analytical mind and SQL skills, so the work fell to the IT manager. Exports and large result sets also slowed down once they passed a certain size.
02
The Solution
An outside consultant pointed Double Wood to Sigma because it could do more than its legacy BI tool, with a simpler interface and the ability to drill down to the SKU level for more granular insights. Double Wood had already built out years of dashboards in that legacy tool, so Sigma’s Migration Skill gave them a way to bring that existing work over with minimal effort. Sigma Migration Skills are open-source tools that direct an AI coding agent to rebuild a BI dashboard in Sigma and verify the result value by value. The skills run on Sigma’s Workbooks as Code API. Double Wood ran them through Claude Code and needed no outside consulting help to do it.
Double Wood had also looked at running its reporting through Claude on its own and found two limits:
Only the person who built a dashboard could refresh it.
A dashboard built for one person could not be shared across the company.
Sigma inherits the permissions and governance already set in Snowflake, so both problems went away when Double Wood switched to Sigma.
03
The Results
Dashboards that would have taken a week or two to convert were completed in just a few hours. The agent did roughly 80% of each rebuild in that window, and IT manager Bill Dempsey spent his time verifying numbers rather than rebuilding layouts. The time to conversion decreased significantly after the first launch as Claude and Workbooks as Code learned conventions while keeping the same banner styles and layout. Bill converted 12 dashboards, several of them 10 to 12 pages long, in about two weeks of part-time work.
The 12 rebuilt dashboards match their BI originals, and hundreds of hours of earlier build work carried over with them. Double Wood’s leadership now have visibility into company health and Amazon settlement, charge, and promotion reporting real-time in Sigma. The team registered its own Sigma sales agent inside the dashboards and asks questions about Amazon, wholesale, and shipped sales against the data in front of them. Sigma runs those queries live on millions of rows in Snowflake, at the speed Snowflake itself returns them.
“It went from a week or two of work to like three or four hours. It’s been pretty incredible.”
Double Wood Supplements results
- Per dashboard conversion
- Weeks to hours
- Dashboards in two weeks
- 12
- Done by the agent before review
- 80%
04
What’s Next
The team plans to increase the amount of data in Snowflake, then expand the dashboards further and put more of them in the hands of the executive team. Double Wood now sees a path to replace at least two legacy applications for demand planning and project management with apps built on Sigma.




