Introduction
CaseGuru is a Russian brand of consumer electronics and small household appliances with direct online sales through its own online store. The company's product range includes headphones, smart watches, vacuum cleaners, speakers, power banks, kitchen appliances, beauty, hygiene, accessories and components. Due to the wide catalog and large flow of requests, support and sales were faced with recurring questions every day: product selection, model comparison, delivery, payment, discounts, order statuses, warranty and returns. To process this flow faster, reduce the load on the team and get economic benefits from automation, for CaseGuru implemented an AI bot connected to the knowledge base and RetailCRM
A large flow of requests began to overload support
Before the launch of the bot, employees manually processed the entire flow of requests: they responded to products, characteristics, delivery, payment, promotions, order statuses, guarantees and returns. With a large volume of messages, this created a constant burden on the team: responses were slow, employees’ time was wasted on the same type of questions, and the quality and speed of the response depended on the workload of the shift. This was especially noticeable in the most widespread scenarios: consultations on goods, questions on orders and statuses, placing, canceling and confirming orders
CaseGuru needed not just a bot for answers, but a tool that would help quickly sort out the flow of requests, automate standard consultations and relieve support and sales staff. At the same time, it was important for the client to maintain the quality of the service: the bot should not invent information, give unconfirmed answers, or lead the client into a false scenario. If there is insufficient data or the situation requires human participation, the request should have been transferred to the operator
We set up an AI bot as the first level of support and sales
The client asks a simple question about the function of the product, and the bot is not limited to a short “yes” or “no”. It explains the model’s features of the model, explains the advantages and continues the conversation so that it is easier for the client to make a purchasing decision
The bot doesn't just list models, and helps you choose the right option. In this example, the bot first suggests available models and then narrows the selection to suit the client's specific request. Due to this, the conversation turns not into a catalog reference, but into a full-fledged consultation, which helps to quickly guide a person to the choice of product
One of the important scenarios is correctly handle the situation when the product is not on sale. In this example, the bot does not interrupt the conversation, but immediately explains the status of the model and takes the client to the next step: offering alternative options and continuing the consultation
When a client writes with a question about an order, the bot does not give a general answer, but immediately accesses the data in RetailCRM and explains the current status. In this example, it shows at what stage the order is, clarifies the delivery address and carefully picks up the emotional request of the client, where timing and urgency are important.
Bot closes requests not only for goods and orders, but also for service. In this conversation, it tells the client what to do even after the end of the warranty period: he explains the procedure for contacting, gives a clear next step and helps not to lose the client on a service request
What did the client receive?
After implementation, CaseGuru received not just faster responses to customers, but a full-fledged first level of automated support and sales
The bot took on standard consultations on products, selection of models, answers on characteristics, orders, delivery, payment, guarantees, returns and components
Employees stopped spending most of their time on repetitive questions and were able to connect only where manual work was really needed. Support has become more stable, and customer experience has improved smoother even with a large flow of conversations
