The international future of customer experience in retail: voicebots and visual search
About the customer
CCC S.A., based in Poland, is one of the largest European footwear and omnichannel retailers. It operates a unique footwear and apparel ecosystem of over 90 ecommerce platforms and 1,000 shops across 29 countries under various brands and channels.
About the project – the challenge
To enhance their customers’ shopping experience, CCC looked for a solution that could make shopping smart, in line with the latest technology and trends.
They wanted to introduce a reliable 24/7 customer service in several markets, facilitate product search for a varied group of clients, and catch up with the latest trends in retail technology on different platforms.
Yosh’s solution included three elements: a voicebot, a chatbot, and a visual search engine.
The voicebot processes human speech and then produces an audio response, which allows for natural, intuitive voice-based communication, without pressing any buttons. By leveraging the advanced functionalities of Google Cloud Dialogflow and Yosh’s expertise, we designed a project that is easy to navigate for the customer and easy to scale for the owner. It is available via phone and Google Assistant for the Polish market.
The chatbot is available in Polish on the CCC website, WhatsApp, Google Assistant, and in six different markets on Facebook: Polish, Austrian, Czech, Slovak, Hungarian and Romanian. Similar to the voicebot, it relies on Google Cloud Dialogflow, but was designed as a separate entity with chat communication in mind. Thanks to Yosh’s linguistic and technological know-how, as well as Dialogflow’s flexible architecture, the product guides the customer within the constraints of the chat medium.
The chatbot and the website are additionally enhanced by state-of-the-art visual search technology. To find the desired product, the client skips linguistic input and simply uploads any photo, whether it was downloaded from another website or snapped on the street. Then, an AI model, trained to recognize similarity between images, returns a list of the most similar products from the store’s offer. All operations are processed on Google Cloud and are available on markets in Poland, Slovakia, Czech Republic, Hungary, Bulgaria, Austria and Romania, with more on the way.
The supported functionalities can serve the whole shopping journey: from simple product search in the chatbot, through visual search, which is particularly important in apparel, to post-purchase support, via order status checks and general FAQs with the voicebot.
Yosh’s solution also resolves issues related to operating internationally: an easy-to-maintain core structure is customizable for a wide range of languages and markets, and the list keeps growing.
Less than a year after the initial implementation of visual search in August 2020, the number of visual product searches has exceeded 1 million, with an average of 150,000 searches per month in 2021.
Within the same timeframe, the voicebot handled over 40,000 minutes of conversations with CCC’s clients. The bot answers all calls outside the working hours of the human consultants, which made the CCC helpline available 24/7 and reduced the workload during peak hours, allowing consultants to focus on more complex issues.
The chatbot solution is being developed with different ranges for different markets, whose list is growing.