Spryker Hackathon – Chatbot Component for the Commerce System
We had a really relaxed start to the Spryker Excite conference. Good coffee, early conversations over breakfast – the usual. But then, at 8.30 am on the dot, the hackathon started. Ruben and I were joined by Victor and Alvin from Jungheinrich in Madrid and Rene from diva-e in Hamburg. The teams were drawn by lot.
Spryker Hackathon, the Task
The above-mentioned Spryker component is, of course, based on Spryker Cloud Commerce OS, the Spryker operating system for commerce applications. For the hackathon, we could choose from ideas that were generated in an online idea session beforehand, or we could contribute our own ideas. And since we had attended an AI workshop with our colleagues from Wroclaw only recently, the topic was fairly obvious.
After coordinating with the team and including the additional requirement to use the new composable storefront framework Oryx, we decided to implement a chatbot. The intention is for it to handle customer requests in the webshop in the future. The necessary information is gathered from existing CMS pages of the particular shop. The ‘large language model’, or LLM for short, which the AI is based on, initially only receives training data from the information in the content management system. The special feature: If the AI cannot provide a customer with a satisfying answer, the component will alert support or the back office to address the specific question. First level support is covered by the chatbot, second level support by a human.
Customers Save Time, the Company Saves Money
Why are these components particularly important in e-commerce? The customer experience in customer support improves while, at the same time, the workload gets lighter. This is also because the chatbot answers a lot of questions right off the bat, which shortens the response time. We also found that many webshops already have FAQ pages. However, these are usually hard to find for customers or distributed across different sections. A chatbot provides this information in an easy-to-consume format that is tailored to the specific customer.
Hackathon, Our Approach
After the networking breakfast at the Spryker office and a short brainstorming phase, we split the project into different segments to make headway faster:
First, we searched for public AI libraries to develop the AI part of the backend.
We then implemented a ‘Glue API interface’ to link the frontend and backend components.
Next was the frontend chat component for the Oryx composable storefront framework.
Of course, we also had to develop a back office component which would enable admin and support users to view and answer customer questions when necessary.
The key to our approach: We split the tasks in the team and helped each other out when anyone needed support.
The hackathon and its result showcase another opportunity to develop components for your own business – tailored increasingly to the individual needs of customers.
High-speed Development for Spryker and Unic Customers?
So, what is the takeaway from this event for us as Unic employees? One big advantage of the Spryker composable commerce platform is the efficient modularity of the commerce system. Many components are available from the Spryker shop right from the start.
The hackathon and its result showcase another opportunity to develop components for your own business – tailored increasingly to the individual needs of customers. The new Oryx framework increases this flexibility in the frontend even more. Without this flexibility, it would have been difficult for us to implement our chatbot component in such a short time – and we are talking one working day here.
So Now What?
The code of the component has been published in the Spryker community repository on GitHub under an MIT licence. This means that the component draft developed by us is available to every developer as a basis for your own chatbot or similar components. To try the AI part, however, you need an OpenAI key.
The entire event was a lot of fun, but it was also quite a challenge. And if you keep in mind that we only met most of our team just before the hackathon, it was even more surprising how well we worked together as a team.
Contact for your Digital Solution with UnicBook an appointment
Are you keen too discuss your digital tasks with us? We would be happy to exchange ideas with you: Jörg Nölke and Gerrit Taaks (from left to right).
Contact for your Digital SolutionBook an appointment
Are you keen to talk about your next project? We will be happy exchange ideas with you: Melanie Klühe, Stefanie Berger, Stephan Handschin and Philippe Surber (clockwise).