OonkoO

BACKGROUND

Please describe your company and position.

I am a Brand Officer at Transcom Beverages Limited

Describe what your company does in a single sentence.

Transcom Beverage Ltd (TBL) is the exclusive PepsiCo Franchisee for Bangladesh. TBL owns and operates modern plants in Dhaka and Chittagong for bottling the renowned soft drink brands such as, Pepsi, 7UP, Mirinda, Slice, Mountain Dew, Pepsi Diet and 7UP Light.

OPPORTUNITY / CHALLENGE

What specific goals or objectives did you hire OonkoO to accomplish?

• Image processing service to audit our visi coolers in the market.

SOLUTION

How did you find OonkoO?

• Online Search
• Referral

Why did you select OonkoO over others?

• Pricing fit our budget
• Good value for cost

How many teammates from OonkoO were assigned to this project?

2-5 Employees

Describe the scope of work in detail. Please include a summary of key deliverables.

I contracted Oonkoo to design and implement a machine learning-powered computer vision system for auditing the purity of visi coolers deployed in the field. As part of our operational workflow, Product Sales Representatives (PSRs) are required to capture images of the visi coolers immediately after completing their sales visits.

The system developed by Oonkoo leverages deep learning—most likely convolutional neural networks (CNNs)—to analyze these images and classify them as either “pure” or “impure” based on predefined visual criteria. The model is trained to detect violations such as the presence of unauthorized products, improper arrangement, branding inconsistencies, or other anomalies that compromise cooler purity.

This automation enables scalable, consistent, and near real-time auditing of field assets, reducing dependency on manual inspection and improving compliance monitoring across our retail network.

RESULTS & FEEDBACK

What were the measurable outcomes from the project that demonstrate progress or success?

The deployment of Oonkoo’s machine-learning audit system yielded the following quantifiable outcomes:

Model Performance

Achieved an overall classification accuracy of 80%.

Precision and recall metrics consistently met or exceeded target thresholds for both “pure” and “impure” cooler classes.

Operational Efficiency

Doubled audit throughput, enabling 2× more audits per given time period.

Reduced the average time per audit by 20 minutes, freeing field resources for additional value-added activities.

Compliance & Corrective Actions

Corrective action completion rate among PSRs rose by 40%, reflecting faster resolution of identified purity violations.

90% of PSRs consistently uploaded post-visit images, demonstrating high user adoption and system reliability.

Business Impact

Cleaner, more compliant cooler stocking led to measurable improvements in brand block visibility on-shelf.

The outlet “VPO” metric increased from 20 to 30, indicating a 50% uplift in effective sales-call outcomes per location. 

Describe their project management. Did they deliver items on time? How did they respond to your needs?

Yes, they were very sincere about the time.

What was your primary form of communication with OonkoO?

• Virtual Meeting
• Email or Messaging App

What did you find most impressive or unique about this company?

Domain-Specialized Vision: Custom datasets and tuned CNNs detect subtle purity issues—misbranding, misplacement, stray products.

Rapid, Scalable Deployment: Live in weeks, doubling audit throughput and cutting 20 minutes per audit via a containerized inference stack.

High Adoption & Impact: 90% PSR participation, 40% higher corrective-action rate, and a 50% uplift in outlet VPO (from 20 to 30).

Are there any areas for improvement or something OonkoO could have done differently?

Show Exactly What’s Wrong: Instead of just “good” or “bad,” highlight the exact shelf area or item that’s out of place.

Clear Manager Dashboard: Add easy-to-read maps and charts so managers can see where problems pop up and focus their efforts.

Explainable Results: Give reps a confidence score or simple visual cue (like a heatmap overlay) so they trust and understand the AI’s decision.