BPO: AI and agility revolutionize menu digitization

A standout software company — a leader in solutions for the service industry — identified a critical opportunity to transform the menu digitization process for restaurants on behalf of their clients, among them Teleperformance, a top business process outsourcing (BPO) provider for customers like Uber Eats.

Client context

Our client's vision was to build an innovative solution that companies like Teleperformance could deploy to optimize their digitization processes. With a forward-looking approach, they wanted to go past the limitations of existing manual systems and give their clients the ability to onboard new restaurants to delivery platforms faster and more efficiently.

The challenge

The challenge was to build a scalable, adaptable solution that could handle diverse menu formats and keep up with the growing demand in the food-delivery market.

Several key opportunity areas surfaced:

• Automate the manual menu transcription process to cut time and errors.
• Boost processing capacity to onboard new restaurants faster.
• Ship a scalable solution to meet the market's growing demand.
• Guarantee consistency and accuracy of the digitized data.

The Centraal solution

Centraal designed and shipped a solution that brought the client's vision to life:

Image Analysis System

• We built a Google Cloud Functions–based backend that integrates machine learning models (OpenCV and Sklearn) to extract information from menu images.
• The solution automatically processed diverse menu formats, regardless of their structure or image quality.

Optical Character Recognition

• We deployed Google Cloud AI's OCR to convert image text into structured data.
• This enabled precise, fast digitization of menu information.

Agile Approach and MVP

• We adopted agile methodology to rapidly build and refine a Minimum Viable Product.
• We ran continuous iterations based on client feedback to optimize the solution.

Centraal Studio designed and shipped a digital marketing campaign that let us validate our product quickly and with controlled investment. Their innovative approach let us make informed decisions and move forward with confidence!

— Tomás Palacio, Founder and Director — Ex Procurement

Collaborative delivery process

We adopted an agile approach to build a Minimum Viable Product (MVP) that tackled the menu digitization challenge. Through Lean methodology workshops, we identified the key requirements and prioritized the essential features.

Results and benefits

Shipping our solution powered the transformation of the digitization process:

• 75% reduction in menu digitization time.
• Significant boost in processing capacity for new restaurants.
• Minimized human error, improving data consistency and accuracy.
• Greater adaptability to respond to growing market demand.
• Improved scalability, enabling fast service expansion.

We built the solution using OpenCV for image analysis and Sklearn for machine learning. The backend ran on Google Cloud Functions, taking advantage of its serverless capabilities. Google Cloud AI OCR handled text recognition, while Angular powered the user interface. We managed the entire process with Jira, ensuring efficient collaboration and iterative development.