As a trusted product tester within the food supply chain, FoodChain ID are constantly striving for more innovative products that make it more efficient to connect and use both external, and internal data. These data-driven solutions help manage entire client portfolios of products, throughout their life cycles, to ensure they remain compliant, transparent and safe.

This need for stringent, compliant and accurate data, meant that FoodChain ID needed to tackle a number of cloud challenges, helping them continue to innovate and lead in their field. These challenges ranged from effective data understanding to cloud storage and classification of product types. Product data ranged from food types, such as vegan, vegetarian and plant-based to GFSI food safety certifications and non-GMO ingredient verification, to name a few.

FoodChain ID had worked with third parties in the past, in order to plan and architect the structures that support their core services. They were looking for a cloud specialist that could help them further advance their initial goals of superior data extraction. As a specialist in machine learning services and Intelligent Document Processing (IDP), within Amazon Web Services (AWS), Firemind were uniquely placed to be a partner that could help FoodChain ID reach their business objectives.

We could also support them with an accelerated project, helping to theorise, architect, build and implement an innovative solution, with support and backing via AWS.

FoodChain ID were going to be using a wide array of AWS services within this partnered project, in order to advance their capabilities and deliver a working end-to-end solution. Services that could very quickly replace old architectures, and turn them into faster, more modernised and well-governed ‘drag and drop’ workflows.

We began with Amazon Textract (an Artificial Intelligence (AI) tool), to automatically extract the text within the PDFs. Textract works cleverly, going beyond simple Optical Character Recognition (OCR), by extracting relationships and defined structures from document types. This helps to reinforce higher accuracy within the labelling results.

Then by setting up Amazon Comprehend, part of the AWS Natural Language Processing (NLP) family of services, we could understand and extract desired insights from text within any document type, perfect for this particular case, where PDF extraction was key.

This is where things get interesting. Rather than simply relying on Comprehend to extract and understand particular values, we used Doccano to drive the retraining data logic, ensuring that there is a corrective learning loop.

Doccano is an open-source annotation tool which provides exceptional text classifications, entity extractions and sequence to sequence translations.

Doccano doubled as both a dataset labeller, and a ‘human’ port of call, for any low confidence scoring from Comprehend. This is useful for the early stages of testing, before deployment, as human operators in both our team and FoodChain ID’s, could easily assess and update the training models – leading to even higher accuracy.

With the introduction of an Intelligent Document Processing workflow, FoodChain ID could refocus their time on product development and enhancements to their customers. Saving heavily on resource costs and time spent manually inputting data fields.

Our solution’s integrated human loop meant that the FoodChain ID team could quickly retrain data models and tweak the outputs for enhanced accuracy over time. Producing a self learning and adaptive document system which can become more precise for all classifications over time.

As an Advanced Partner of Amazon Web Services (AWS), Firemind was in a unique position to enable funding opportunities and avenues of cost savings, subject to certain criteria. This guidance and support becomes a valuable asset, when planning, experimenting and building new automated workflows in the cloud.