We’ve enabled a leading global freight-forwarder to automate the e-mail routing and data capture process, saving 42,000 man hours annually for their offshore back-office team
What we were solving for
Reduced workload: Given the large volume of customer queries received daily, dedicating resources to classification/routing of emails or summarising reduces the resources available for tasks that actually create value for the Customer
Scalability: A similar solution is now being developed for other business functions, as it automates the flow of unstructured data to source system, reducing the cost/task materially
OptiSol solution
We have built a machine-learning (ML) based classifier and a Power Automate bot to automate this process end-to-end
We built two levels of ML classifiers, and trained the model with large volumes of historical classification
We used a custom NLP pipeline to identity and extract relevant entities like tracking number, shipment details etc., to ease subsequent review of the emails
Various solutions including OCR solutions have been applied to capture data from PDFs, excel documents and scrapping websites