Case-Study-AI-Models-for-Predicting-Inventory-Restocks-on-Blinkit-&-Zepto

Overview

In the fast-paced world of quick commerce (Q-commerce), efficient inventory management is vital. Blinkit and Zepto, two leading players in India’s 10-minute grocery delivery ecosystem, face constant pressure to keep shelves stocked without overburdening storage. Actowiz Solutions deployed advanced AI models backed by real-time web scraping and historical pattern analysis to accurately forecast inventory restocks—ensuring zero missed sales and reduced wastage.

Challenges

Before partnering with Actowiz Solutions, both retailers faced:

Stockouts during high-demand periods, leading to revenue losses.
Excess inventory of non-moving SKUs due to inaccurate demand prediction.
Lag in vendor restock alerts, especially in pin code-based delivery zones.
Lack of SKU-level granularity in forecasting replenishment windows.
Manual coordination with warehouse ERP for perishable goods.

Objective

Predict restocking needs of high-frequency SKUs across cities.
Build a zone-wise AI model trained on pricing, order rate, and stock visibility.
Reduce product unavailability alerts on Blinkit and Zepto apps by 70%.
Improve vendor coordination based on near-term demand patterns.

Actowiz’s Approach

1. Real-Time Scraping Engine

Using Blinkit and Zepto's front-end interfaces, Actowiz set up a high-frequency scraping system that extracted:

Product availability by pin code
Stock status (e.g., "Only 2 Left", "Out of Stock")
Price changes and discounts
Reappearance timestamp of previously out-of-stock SKUs
Location-wise category rank

2. Historical Dataset Modeling

Actowiz gathered 6 months of SKU availability snapshots (every 30 minutes) for key cities like:

Delhi NCR
Mumbai
Bengaluru
Pune
Hyderabad

Sample data structure:

Date	City	SKU	In Stock	Quantity Tag	Reappeared After (hours)
2025-05-12	Mumbai	Amul Toned Milk 1L	No	-	4
2025-05-12	Mumbai	Amul Toned Milk 1L	Yes	"Only 3 left"	-

This data was used to train time-series models for each product + pin code combination.

3. Machine Learning Models Used

LSTM Neural Networks for SKU restock time prediction
Random Forest Classifiers for binary classification (restock likely/not likely in next 12 hours)
Gradient Boosted Trees for multi-variate regression on stockout duration
Bayesian Inference for estimating restocking frequency during promotions

4. Integrating External Signals

The AI models were enhanced with:

Blinkit & Zepto app promotions
Festival dates & traffic spikes (e.g., Diwali, Independence Day)
Weather data for fresh produce perishability
Vendor-wise replenishment cadence

Key Features of Actowiz’s AI System

Feature	Description
SKU-Level Prediction	Hourly probability of restock for individual SKUs by city
Stockout Alert System	Dashboard alert for high-demand SKUs nearing depletion
Price vs. Restock Analysis	Predict restocking lag based on recent discounts or flash sales
Vendor Mapping	Correlate restock speed with past vendor delivery times
City-wise Heatmaps	Visualize restocking rates across hyperlocal zones

Dashboard Snapshot (Sample)

A dynamic dashboard powered by Actowiz displayed:

For Blinkit – Andheri West, Mumbai

Top 10 SKUs predicted to be out-of-stock in next 4 hours
Estimated replenishment time (in hrs)
Historical avg. restock window

SKU Name	Stock Status	ETA to Replenish	Demand Surge (%)
Tropicana Orange 1L	Low (Only 2)	3.5 hours	+42%
Britannia Bread 400g	Out of Stock	5.2 hours	+37%
Dettol Handwash 250ml	Low	2.0 hours	+18%

Business Impact

Operational Wins:

Metric	Before Actowiz	After 90 Days
Average Stockout Rate	17.3%	4.5%
SKU-Level Forecasting Accuracy	62%	89%
Reduction in Missed Orders	-	31% increase
Vendor Coordination Delay	8 hours avg.	2.5 hours
Manual Alerts	100+ daily	Auto AI-based

Client Testimonial

"Actowiz’s predictive inventory system transformed how we plan replenishments. We no longer chase restocks manually. Instead, we’re proactively ready for every demand surge."

– Inventory Head, Blinkit (Mumbai Ops)

Expansion Plan

Actowiz is now extending this AI restocking solution to:

Regional warehouses of Zepto in Tier 2 cities like Indore, Kochi, Lucknow
Integration with dark store partners
Predictive insights for perishable SKUs with <24hr shelf life
Zepto Café’s inventory for combo meals & snacks

Conclusion

This case highlights how real-time data scraping, when paired with AI/ML modeling, can unlock powerful restocking predictions in hyperlocal delivery models. For Blinkit and Zepto, Actowiz Solutions helped move from reactive inventory handling to predictive precision—ensuring consumers always find what they need when they want it.

Real Results from Real Clients

Hear It Directly from Our Clients

Watch how businesses like yours are using Actowiz data to drive growth.

▶

1 min

★★★★★

"Actowiz Solutions offered exceptional support with transparency and guidance throughout. Anna and Saga made the process easy for a non-technical user like me. Great service, fair pricing!"

Thomas Galido

Co-Founder / Head of Product at Upright Data Inc.

▶

2 min

★★★★★

"Actowiz delivered impeccable results for our company. Their team ensured data accuracy and on-time delivery. The competitive intelligence completely transformed our pricing strategy."

Iulen Ibanez

CEO / Datacy.es

▶

1:30

★★★★★

"What impressed me most was the speed — we went from requirement to production data in under 48 hours. The API integration was seamless and the support team is always responsive."

Febbin Chacko

-Fin, Small Business Owner