Turning a 24-hour AI hackathon into an operational decision tool.
PromoPulse is an AI-powered promotional readiness concept created during the 2025 AI Hackathon to help retail teams identify stock risk and competitive pricing threats before campaigns go live.
Note on data: All values shown are illustrative and do not represent actual company data.
Store managers often receive promotional plans without clear visibility into local inventory readiness. That gap drives stockouts, reactive transfers, and poor customer experience. At the same time, competitor price changes are hard to monitor at scale, leaving teams without timely signals to protect margin and demand.
Stakeholder Ecosystem
The product supports a multi-tier retail hierarchy, from frontline store decisions to regional intervention planning.
The 24-hour window ruled out designing a broad autonomous promotion platform. I narrowed the concept to one defensible intelligence flow: ingest competitor circulars, use OCR and LLM pipelines to structure pricing signals, match them to internal SKUs, and surface early warnings before a campaign launches.
Showing every AI-detected difference would have recreated the overload store teams already faced. I designed the system to suppress routine items and prioritize only high-risk products across thousands of SKUs, turning model output into a focused intervention queue.
Flags products using risk signals such as low inventory, historical underperformance, and adverse price position.
Visualizes price differences against key competitors, including percent deltas that inform tactical response.
The interface embeds live Regional Distribution Center telemetry directly in the decision workflow, giving teams supply context at the moment they review risk.
A visual timeline tracks fulfillment from order placement through RDC processing to delivery with ETA guidance.
Granular controls let users route alerts by channel and category, including pricing and inventory events.
Moved from a broad AI opportunity space to one prioritized, high-fidelity promotional-readiness workflow within the hackathon window.
Designed the risk engine to surface only high-risk promotional items, reducing a large product set to an actionable review queue.
Connected store-level intervention with regional oversight so managers could move from portfolio risk to location-specific action.