Automation & Scaling

Business Levers

Fleet Profiles
Demographics & Time zones
Delivery Volumes

Key Dimensions

Customer
Geo
ML Derived Inputs
  • Router Mux
  • Truck Driver Segments
  • Delivery Volume Patterns
Constraints
  • Average Delivery Time
  • Number Of Stores Served
  • Budgetary Limits
Optimizer
Recommendations
  • 8 Mid Volume Trucks
  • 32 Down Town Stores
  • 6 Experienced Drivers
  • 20 Mile Radius
Historic Usage And Supply Trends
  • Forecast inputs
  • Sales inputs
  • Supplier constraints
  • Workforce challenges
External Factors
  • COVID-19 data
  • Macro economical
  • Manufacturing challenges
  • Import/Export policies
Temporal Features
  • Fiscal patterns
  • impliSeason and weather cations
  • Demand fluctuations
ML-Based
Prediction
Model
Prediction Scorecard
  • Item 1 – 80%
  • Item 2 – 70%
  • Item 3 – 40%
  • Item 4 – 20%
  • Item 5 – 60%

Business Challenge

Customer has a large manual process to examine and identify containers with physical damages. The existing process was tedious and time-consuming, prone to manual errors while tracking defects.

Our Solution

We helped develop a deep learning-based video processing solution which could identify the defects on the containers and automatically log the container details into the existing data system.

Outcome

The solution has been successfully tested and is currently in the operationalization phase.

Business Challenge

Catering to multiple clients across the globe, the call center company was dealing with the challenge of unoptimized agent utilization.

Our Solution

We developed an ML-based solution which identified and recommended the ideal agent profile for a particular shift after analyzing all the demographic and firmographic factors of the company’s clients.

Outcome

Our recommendations resulted a significant reduction in MTTR by up to 12%.

Business Challenge

The leading e-commerce company offers a wide mix of deals to its customers across multiple categories and locations. However, the product assortment is not optimized well for demand, leading to frequent inventory problems.

Our Solution

We analyzed the situation and designed a system that considers location, seasonality and nature of deal to identify an optimal assortment of products. It can also be tuned further to improve topline or bottom-line traffic.

Outcome

The recommended deals helped the company enhance their revenue by up to 15%.

Business Challenge

The company was unable to derive value from its digital marketing initiatives due to heuristic lead scoring and visitor engagement models.

Our Solution

We developed a statistical model to establish a data-driven model for lead scoring. We also built ML-based channel attribution models to measure the efficiency of digital content.

Outcome

The percentage of conversions through digital channels has increased from 4% to 7% after implementing the recommendations provided by our ML-based models.

Business Challenge

There are close to 2.5 million users who have registered with this e-commerce platform. However, even a single purchase had never been recorded due to lack of ability to identify which is the right user segment that can be targeted to increase engagement.

Our Solution

Based on the browsing pattern of the new users and the existing users, we developed a machine learning model that could predict and identify prospects who have a high propensity to make a purchase.

Outcome

The enterprise recorded up to 12% increase in conversions.

Business Challenge

The company designs and deploys custom refrigerators at various supermarkets. However, it did not have a mechanism to figure out when the refrigerators would fail.

Our Solution

We designed an ML-based predictive model which can predict the probability of a refrigerator experiencing downtime by using the sensor data collected by IoT devices deployed in a refrigerator.

Outcome

Our recommendations resulted a significant reduction in MTTR by up to 12%.

Business Challenge

Predict the likelihood of farmers discontinuing the engagement with agricultural company.

Approach
Business Challenge

Provide an optimized prescription on product usage to farmers to get the best yield.

Approach
Business Challenge

Understand visitor behaviour and profile the traffic landing on the agricultural company’s website.

Approach

A $40 Bn company which is into manufacturing networking hardware, software, telecommunications equipment and other high-technology services and products.

Business Challenge

Sales planning teams do not have a driven perspective of “Account Potential” for effective forecasting and quota planning, and do not have visibility into the health of “Opportunity” pipeline.

Our Solution

We developed an ML-based opportunity health scoring model built out of GCP to flag the opportunities as healthy/unhealthy and integrated results in sales dashboards. Provided an ML-based recommendation engine built upon the existing product mix data to furnish persona-based recommendations based on propensity of purchases.

Outcome

Performed key driver analysis to understand the major contributors for conversions. Transformed sales planning based on analytical insights and helped record an increase of 1.7x in conversions.

Overlay the Traditional Whitespace With Product Affinity for Hyper Targeting of Customers

How: Fully Operational Dashboard

Who: Sales and Marketing Teams for Planning and Promotion

Product Affinity – ML
Business Review and Approval

Overlay the Traditional Whitespace With Product Affinity for Hyper Targeting of Customers

DIGITAL NATIVES

This group of farmers is highly digital-savvy and tend to use the latest tools and trends for farming. We can position digital solutions for this group.

FARMING IMMIGRANTS

This group of farmers is new to farming digitally and might need special customer experience tailored for their needs.

LOYAL CUSTOMERS

This group of farmers has been consistently purchasing products. We may look at upselling more to them or design a customer rewards program for them.

ONE-TIME EXPLORERS

This group of farmers has a high tendency to not return to make a purchase. Identifying then ahead of time and engaging them well in advance would help them to stay connected and return to make a purchase.

AI-Based Forecasting System to Predict the Demand Accurately for Operational Efficiencies

Historical Data
  • Sales
  • Promotions
  • Digital Touchpoints
External Data
  • Weather
  • Economics
  • Competitors
AI- Based Solution
Top Down Forecast

Demand predictions at a country and product level.

Bottom Up Forecast

Demand predictions at the SKU and lowest level of geography (city/county level).

Discovery Insights
  • High Demand Areas
  • Low Supply Areas
  • Low Density by farmers
  • Low Market Capture Zones

Automated service model to generate maximum value from BI assets

Data to Decisions

Existing BI Landscape

Decision Boards to Self-Service Portals

Self-Service Dashboarding Capability
KPI Bucket

Repository of KPIs across all the BI reports & ADS irrespective of technology & tool type

KPI Viewer
  • UI to Visualize KPI of interest
  • Use KPIs to create decision boards on the go
  • Role-based access based on user personas

Automated workflow to generate maximum value from BI assets

Baseline Goals
Design Data/Biz Rules Fabric
Build The Experience

Establishing technical, operational and performance baselines and defining the deliverables to accomplish the project objectives.

Modelling dashboards with the right KPIs & business rules to create analytical insights that are critical for business outcomes.

Creating dashboards that engage the audience to make evidence-based decisions and make the reporting more appealing, self-explanatory, and memorable.

Critical Success Factors
  • To ensure maximal knowledge transfer in both ways we would need to work closely with key experts in client’s business and IT departments.
  • Rapid access to potentially disparate data source and support in understanding the data is essential in order to build up the data structures required for the analytical models.