Cisco eCommerce: Building AI powered chatBOTs, leveraging deep learning algorithms and OCR

Use cases for leveraging AI have been well documented for consumer facing eCommerce platforms but there has been less progress in incorporating AI in large B2B eCommerce platforms with very complex high tech products. Cisco’s eCommerce platform is responsible for buying & selling all of Cisco’s products, including hardware, software, services & subscriptions. It is a custom-built mission critical platform which accounts for $40+ billion of Cisco’s revenue annually. Cisco is under going a transformation to move it’s hardware revenue as recurring subscriptions based revenue. Cisco is also looking at additional avenues for increasing revenue by providing additional platforms for Direct/SMB customers who are not well versed with how to order Cisco’s complex products. This presents a number of challenges for our users to understand the ever-changing user experience. We have been leveraging AI & Machine Learning to provide a guided experience to our users to bring them onboard in this transformational journey. We have built multi-lingual AI powered chatBOTs leveraging Google’s Cloud platform. We have also leveraged machine learning algorithms to find popular products by category, items sold together and products offering best discounts which can change dynamically. BOT also has the capability to understand the sentiments of the user & route it to a live agent if user is unhappy. We have built a deep learning (TensorFlow) & OCR based algorithm to automate Fax orders. Cisco processes these orders manually as the order format is different for different partners & can change dynamically. By automating, the processing of these orders will result in freeing up 40+ customer support resources. Business KPIs being tracked are:- 1. Cost Reduction 2. Customer Satisfaction (CSAT) 3. Automation (Self Service) 4. Cycle time 5. Additional Revenue due to upsell