VendorIQ is a Power BI dashboard I built from the ground up to solve a real problem: reliability across a 150+ vendor network lived entirely in people's heads, with no consistent record of who delivered on time and who didn't. Using three years of historical delivery data that had never been centralized, I logged metrics by hand where no record existed, then layered in direct connections to internal systems as the pipeline matured, so VendorIQ now runs on a mix of manually imported and live-connected data. It replaced institutional memory with a documented, queryable performance record, and it functions as a real decision-making tool, not a static report. When a vendor delay threatened a hard project deadline, VendorIQ surfaced an alternate, expedite-capable vendor in time to hit it, a decision that used to depend on who happened to remember which vendors were reliable. This was built for internal use, not as a public product, so this page is a description rather than a link to a live demo.
Platform
- Built and maintained in Power BI
- Data sourced from a mix of manual imports and direct connections to internal systems
- Three years of historical delivery data logged by hand to establish the baseline
- Integrates with the Anthropic API to auto-grade new vendors on onboarding
How It Works
- VendorIQ turns three years of historical vendor data into actionable insight on which vendor to choose for a specific project
- Vendor performance is represented as an overall health rating in the dashboard
- Sortable by lead time accuracy, quality of product, responsiveness, and other metrics
- Searchable across the full vendor network by product or material
- Vendors are graded on an A to F system across each metric
For example, if I needed a specific material at a specific location by a specific date and price, I can sort and find the best vendor to source from while seeing exactly how accurate they have historically been at delivering, so I choose the optimal vendor for every project, every time.
AI-Assisted Grading
- When a new vendor is added, VendorIQ calls the Anthropic API to research it automatically
- Reads customer reviews and public sentiment to estimate a starting grade across every tracked metric before a single order has been placed
- As real procurement data comes in, actual performance gradually averages into and overrides that initial estimate
- Grades self-correct over time based on logged experience rather than staying fixed at the AI's first estimate
For example, when a new vendor is added, VendorIQ scours the internet, reads customer reviews, and gauges public sentiment to estimate a starting rating across every metric it tracks, before a single order has been placed. As real purchases are made and performance is logged, quality, and lead time accuracy, those initial estimates slowly update and average based on actual experience.
Scope
- 150+ vendor network
- 3,000+ SKU catalog
Impact
- Replaced institutional memory with a documented, queryable performance record
- Used live to resolve a client-facing vendor delay by surfacing an expedite-capable alternative in time to hit the deadline