Client Introduction

Based in Paris, our client is a well-known global IT startup. With 110 employees, our client digitalis’s retail execution and merchandising for manufacturers and retailers globally. Their AIbased solutions are helping many multinationals across the globe. The client needed Image Annotation (Product recognition services) to support their Artificial Intelligence-based solutions.

Requirements & Specifications

As the client aims to digitalise retail performance and merchandising, they required accurate and customised image annotation services (product recognition services) that can instantly analyze and recognize millions of products and provide quick useful retail insights for efficient retail management of various supermarkets, stores and malls. The client needed a five steps image annotation/product recognition process that included actions like facing, labelling, recopy, pricing, rearranging and verification of each image. For this, the client required a data entry outsourcing partner who could provide Image annotation services (product recognition services) for a massive image volume accurately, without sacrificing quality.

Solutions

After the completion of formalities such as SLA & NDAs- the work began as the client anticipated. We developed a training module in which our in-house
trainer guided our image annotation/ product recognition experts regarding specific client requirements to ensure smooth and successful project execution.

Teams, project Coordinators, Team Leaders, QC Experts Assigned and the Project Begins

A team of 125 image annotation (product recognition) experts were assigned to the project along with five team leaders, one project coordinator and one quality control lead. The team owned excellent image annotation (product recognition) skills and was capable of handling large volumes of projects.

Process

• We began by accessing the image database in the client portal.
• Each image was recognized and boxes were created around the images.
• Every product was properly annotated (with tags and keywords)
• Quality checks were conducted on all images. As the output match the quality standards, it was handed over to the client.

Challenges faced by us

• The complete process of facing, labelling, recopy, pricing, rearranging and verification of each image must be concluded within 5 minutes.
• Given the strict deadline, conducting the procedure on 15,000 to 20,000 images required accurate quality checks, long working hours and strict time management.
• The project required significant attention to detail. However, we ensured that we lived up to the client’s expectations and provided them with accurate data.

Work volume, quality, accuracy & turnaround from our end

Aiming for a satisfactory and useful outcome, we accommodated the client’s requirements by extending our working hours. Daily we provided Image annotation (Product recognition services) of 15,000 to 20,000 images as expected by the client. With proper quality checks, we maintained a 99% accuracy level.

Conclusion

Complete client satisfaction, leading to the establishment of a long-term relationship: the initial 3-month contract has now been extended to a continuous working relationship and outsourcing of more projects. A happy client is our true gain here. The project went on to become a long-term relationship. The initial 3-month contract extended to steady working relationships, outsourcing of more projects and on boarding of more than 100 experts.