E-commerce
Monitor MAP violations across sellers
In this example, we built an agent that uses a Nimble Web Search Agent to search the open web for sellers of protected SKUs and flag listings priced below Minimum Advertised Price with cited evidence.
Quick Start
Inputs
- Protected SKUs The brand's products to monitor, for example "Olaplex No. 3 Hair Perfector 100ml", given as brand, product, and size.
- MAP price The Minimum Advertised Price per SKU, for example "$30.00"; anything advertised below it counts as a violation.
Outputs
- Every online seller of each SKU, with a cited advertised price
- Below-MAP violations flagged by severity
- A repeat-offender leaderboard ranked by seller domain
- A Google Sheets enforcement tracker with one row per violation
- Google Doc notices for the most severe cases, with cited evidence
How it works
A five-stage pipeline.
- Discover every seller For each SKU, a Web Search Agent searches the open web for every seller offering the exact product, capturing the advertised price with a cited excerpt and source URL. Category-group sources steer it toward marketplaces, independent stores, and discounters without restricting it to a fixed list.
- Detect the violations The app parses each advertised price and flags anything below the SKU's MAP in plain code, so the comparison is auditable rather than a model judgment.
- Rank repeat offenders Violations roll up by seller domain into a leaderboard showing which sellers undercut the most SKUs and by how much.
- File and notify The full violation list writes to a Google Sheets tracker; the most severe cases each get a Google Doc notice with the seller, listing URL, price versus MAP, and cited evidence.
- Investigate the evidence A Streamlit dashboard leads with the totals, the leaderboard, and a per-violation drill-down to the exact price excerpt behind each claim.
Stack
Nimble primitives plus the full runtime stack.
Nimble APIs
What it does
- Web Search Agent One dataset_building agent per SKU discovers every seller offering the exact product across the open web and captures each advertised price with a cited source.
3rd Party Tools
Role
- Python 3.10+ Resumable concurrent discovery and the deterministic below-MAP math.
- Google Sheets and Docs The enforcement tracker and per-violation notices, with a local file fallback when Google is not configured.
- Streamlit and Altair The dashboard with severity-coded views, the leaderboard, and the evidence drill-down.
- SQLite The offers and violations store.
Reach out if you have any questions.
Talk to an Expert