222 · 2,681 lines
Invoices processed
Fleet consumption analysis and asset resolution
AI & agentic workflows Internal work sample
Connected consumption records to the correct assets across 222 invoices and 2,681 lines. Resolved inconsistent identities and flagged 24 codes with unusual prices for review.
-
The problem
Inconsistent asset names and aliases obscured consumption patterns and price anomalies.
-
My role
Resolved identities through exact matching, alias normalisation and grounded model lookup for the residual.
-
Where it stands
Processed 222 invoices and 2,681 lines, flagged 24 item codes as price outliers, and delivered a fourteen-tab consumption view.
Results and evidence
24 flagged
Price-outlier codes
14-tab BI
Reporting layer
Business impact
Gives finance consistent asset-level visibility and surfaces price anomalies hidden by flat reporting.
How I did it
- 01
Flagged price outliers against the modal price for each item code.
- 02
Loaded results into a fourteen-tab consumption reporting layer.
Tools and working context
Pairs with the purchase-side invoice pipeline: stock bought and determined there is consumed here, closing the loop from value creation to value recording and evaluation. High-volume consumption invoices matched against a 243-asset reference register.
- Tools
- Python · Gemini (structured output + search grounding) · openpyxl
- Business area
- Consumption analysis · Supply chain
- Skill type
- AI · Data & Reporting
Why I made these choices
- Ordered identity resolution from deterministic matching to grounded model lookup.
- Flagged anomalies for human review instead of correcting them automatically.
What I learned
- Tiered resolution controls cost while reserving grounded lookup for genuine ambiguity.
What comes next
- Feed resolved identities into the reference register and track precision by resolution tier.