02 / AI PRODUCT / VIVRITI NEXT
Less spreadsheet work. More answers.
An AI-assisted workflow that turns Excel inputs into usable dashboards.
- MY ROLE
- Product delivery, from discovery to adoption
- STATUS
- In production · 40+ runs
- COLLABORATION
- Built with the engineering team at Vivriti
The problem
A spreadsheet question could become a request to the data team. The AI Platform gives teams a way to select their data and produce a dashboard through a repeatable workflow.
What I owned
I owned the Excel-to-dashboard capability from discovery and the data-selection contract through processing design, dashboard UX, user acceptance testing, release and adoption.
- 01Select workbooks
- 02Plan the processing
- 03Extract with Python
- 04Build & refresh dashboard
Two decisions
Give the model a bounded job.
Python selects and extracts the actual Excel data. The language model sees workbook structure and proposes a processing plan. This separates flexible planning from deterministic data handling.
Design for the next run, too.
Manual and scheduled SharePoint refreshes retrieve the latest files and update the same SSO-protected dashboard. The product needed to support a recurring task, beyond the first successful output.
What came of it
The capability has been used in more than 40 production runs across teams.
What matters
The product work was as much about defining the data contract and the repeatable experience as it was about the AI.
Public project summary. The workflow diagram is a conceptual overview.
Next: Enterprise AI adoption