BPO: AI and Agility Revolutionize Menu Digitization
A leading software company, known for its industry-specific solutions in the service sector, identified a critical opportunity to evolve the menu digitization process for its clients, including Teleperformance, a global leader in business process outsourcing (BPO) services for clients like Uber Eats.
Client Context
Our client’s vision was to develop an innovative solution that could be implemented by companies like Teleperformance to optimize their menu digitization processes. With a visionary approach, they sought to overcome the limitations of existing manual systems and provide their clients with the ability to onboard new restaurants onto delivery platforms more quickly and efficiently.
The Challenge
The challenge was to create a scalable and adaptable solution capable of handling the diversity of menu formats and meeting the growing demand in the food delivery market. This solution needed to be robust enough to be used by large data processors like Teleperformance, yet flexible enough to adapt to the needs of various clients in the food service industry.
Several key areas of opportunity were identified to drive efficiency and service growth:
- Automate the manual process of menu transcription to reduce time and errors.
- Improve processing capacity to onboard new restaurants more quickly.
- Implement a scalable solution to meet the growing market demand.
- Ensure consistency and accuracy of the digitized data.
Centraal’s Solution
- Image Analysis System:
- We developed a backend using Google Cloud Functions, integrating machine learning models (OpenCV and Sklearn) to extract information from menu images.
- This solution enabled automatic processing of various menu formats, regardless of their structure or image quality.
- Optical Character Recognition (OCR):
- We implemented Google Cloud AI OCR to convert the text in the images into structured data.
- This allowed for precise and rapid digitization of menu information.
- Agile Approach and MVP:
- We adopted an agile methodology to quickly develop and refine a Minimum Viable Product (MVP).
- Continuous iterations based on client feedback were carried out to optimize the solution.
Centraal Studio designed and implemented a digital marketing campaign that allowed us to validate our product quickly and with controlled investment. Their innovative approach enabled us to make informed decisions and move forward with confidence!
Collaborative Implementation Process
An agile approach was adopted to develop a Minimum Viable Product (MVP) that addressed the menu digitization problem. Through Lean methodology workshops, key requirements were identified, and essential functionalities were prioritized. A product backlog with actionable tasks was established for the development team, which included machine learning experts, software architects, and an agile coach.
Results and Benefits
The implementation of our solution drove the transformation of Teleperformance’s menu digitization process:
- 75% reduction in menu digitization time.
- Significant increase in the processing capacity for new restaurants.
- Minimization of human errors, improving data consistency and accuracy.
- Enhanced adaptability to meet the growing market demand.
- Improved scalability, allowing Teleperformance to quickly expand its service.
We implemented a solution using OpenCV for image analysis and Sklearn for machine learning. The backend was developed with Google Cloud Functions, leveraging its serverless capabilities. OCR de Google Cloud AI was used for text recognition, while Angular was employed to create a user interface. The entire process was managed with Jiraensuring efficient collaboration and iterative development.