AskSpot Implementation at 4FIZJO - chat&ticketing system Agent

‍About 4FIZJO

4FIZJO is a dynamically growing e-commerce company founded in 2017 by physical therapists with a passion for sports and rehabilitation equipment. In a short time, 4FIZJO has become a leading player in the fitness and physical therapy market. It currently handles around 120,000 orders per month, employs over 120 people, and has warehouses with a total area of approximately 14,000 m². The company sells its products through its own channels (online stores in Polish and Czech) and through external marketplaces, including Polish platforms (e.g., Allegro) and foreign e-commerce services. This multi-channel sales approach means that 4FIZJO's customer service team must handle a huge number of inquiries coming from various sources. This amounts to thousands of customer cases per month, which 4FIZJO aggregates in a single helpdesk system (Thulium), integrating emails, website contact forms, and marketplace messages (e.g., Allegro Message Center) into one inbox.

The challenges for the 4FIZJO service team were significant. In addition to typical post-sale issues (order information, returns, complaints), customers often asked difficult, specialized questions about sports and physical therapy equipment. Providing advice and helping select the right product for a customer's individual needs was a key element of customer service, requiring expert knowledge and consuming a lot of consultants' time. As a result, with the growing scale of the business, maintaining high quality and speed of service became increasingly difficult. The company recognized the need to automate customer service using AI to offload the team, speed up response times, and simultaneously provide customers with high-level advisory services 24/7.

Goals of the AskSpot Implementation

The implementation of the AskSpot platform as an AI agent for e-commerce at 4FIZJO was of key strategic importance. The following main goals were formulated for the AI-based customer service project:

• Reduce by 50% the number of customer cases and inquiries that need to be handled by a human (both by phone and in writing). The goal was for the AI to take over half of the inquiries coming from all channels, thus freeing up consultants from routine issues.
• Increase the availability of advisory and sales support by providing customers with the ability to get answers to product questions and purchasing advice instantly, 24/7, regardless of the time of day or contact channel. This meant that even outside of customer service office hours, customers could receive expert help.
• Improve the quality of customer service by:
  • Allowing most issues to be resolved instantly on chat, at any time (24/7 self-service without waiting for a consultant's response).
  • Reducing the number of inquiries reaching 4FIZJO specialists and consultants, which allowed them to focus on more complex cases. Unburdened consultants were able to handle the inquiries that the AI could not resolve on its own more quickly and efficiently, which shortened the average response time and increased customer satisfaction.

Stages of AskSpot Implementation

The implementation of AskSpot at 4FIZJO was divided into two main stages, allowing for the gradual expansion of service automation:

1. Stage 1 – Launch of the AI chatbot on the 4fizjo.pl, 4fizjo.cz, and 4fizjo.ro websites. First, AskSpot Chat AI was implemented on the store's websites (Polish, Czech, and Romanian), integrating it with the e-commerce front-end. The chat was configured to handle multiple languages and was made available to customers 24 hours a day. Its capabilities included product search and purchasing advice (the AI can understand the customer's needs and suggest the right product), answering detailed product questions (specifications, compatibility, use, etc.), as well as handling basic order-related issues. Through the chat, a customer can independently check their order status, get delivery information, modify order details, or start the return/complaint process—all without employee intervention, thanks to the AI's integration with the store's system. Moreover, AskSpot on the chat allows for submitting a complaint or return directly in the conversation window, including the option to attach photos/documents. The chat guides the customer through an interactive form, collects the necessary information and attachments, and then transfers it in a structured format to the service system (email/CRM). This solution significantly simplified the complaint procedure for the customer and relieved the service department from manually processing submissions. By implementing AskSpot Chat, 4FIZJO offered a virtual salesperson and consultant on its websites who responds immediately to questions and can resolve most typical issues (from product questions to basic order handling) within a single conversation.
2. Stage 2 – Implementation of the Inbox Agent AI in the helpdesk inbox (Thulium). Following the success of the on-site chat, the operation of AskSpot was extended to passive channels by integrating the AI agent with the internal customer service inbox (Thulium system). The Inbox Agent AI acts as an intelligent assistant "sitting" in the helpdesk: it automatically retrieves new tickets (customer messages) via API, analyzes their content, and recognizes the intent of the inquiry (e.g., product question, request for order status, data change, complaint, etc.). Based on this, the agent can take a series of actions: assign the case to the appropriate queue or department, assign a category and priority, and even perform specific actions in external systems—in line with 4FIZJO's business guidelines. AskSpot was integrated with the store's key back-office systems (product database, ERP/OMS for order management, complaint system, etc.), which allowed it to automatically retrieve the necessary data. For example, the agent can independently check the order status in the store's ERP or the shipment status, verify information about an order placed on Allegro, or extract details about a product the customer is asking about from the knowledge base/specifications. It then generates a draft response to the customer (or a full response immediately) and takes additional actions—such as changing the ticket status, adding tags, or changing the queue.

In practice, the Inbox Agent automatically responds to a large portion of customer messages (from email and marketplace platforms) and prepares responses for consultants in more complex cases, significantly speeding up their work. It can be said that the 4FIZJO inbox has become a living organism in which AskSpot has taken on the role of a virtual employee—sorting and resolving inquiries—which achieves two key goals: ensuring efficient customer service and maximizing sales opportunities with every contact.

Effects of the Implementation

The implementation of AskSpot brought tangible benefits to 4FIZJO, both in reducing service costs and improving customer experiences:

• Reduction in service inquiries thanks to the chatbot: Launching AskSpot on the 4fizjo.pl and 4fizjo.cz websites resulted in a reduction of about 40% in the number of phone calls and messages from customers originating from the websites. Importantly, this reduction in cases did not come at the expense of customer satisfaction; on the contrary, service quality ratings remained high thanks to the fast and accurate responses from the 24/7 chatbot.
• Automation of email and marketplace ticket handling: Introducing AskSpot to the inbox (Stage 2) brought further savings. The Inbox AI Agent independently took over the handling of up to ~33% of inquiries coming from email and sales platforms. In other words, every third inquiry that previously had to be manually handled by a 4FIZJO consultant is now automatically resolved by the AI. AskSpot analyzes messages and provides answers to customers without human intervention in a huge number of typical cases, significantly reducing the number of tickets passed on to consultants. As a result, the actual workload of the service team decreased by another several dozen percent. The remaining (more complex) cases that require human intervention can now be handled more quickly and accurately because consultants have more time per customer. In total, thanks to two levels of automation (chat + inbox), 4FIZJO managed to handle a significantly larger scale of orders without a proportional expansion of the customer service team, which translated into a lower cost-to-serve per order.

Keys to Success

The success of this AI project was due to several key factors:

• State-of-the-art technology tailored to the store's needs. AskSpot was based on advanced, latest-generation large language models (LLMs), ensuring a high level of natural language and context understanding. The use of Microsoft Azure OpenAI and Gemini Cloud guarantees access to leading AI models while maintaining the highest standards of security and GDPR compliance. Equally important was the AI's integration with 4FIZJO's systems—the agent was connected to the product database, the ERP/OMS system, the Allegro platform, and other internal tools. Thanks to an open approach, the solution was configured strictly according to the store's processes and needs. As a result, AskSpot acts as a "virtual employee" tightly integrated into the 4FIZJO ecosystem, understanding the specifics of the company's offerings and procedures.
• Deep understanding of processes and close team collaboration. The close cooperation between the AskSpot implementation team and the 4FIZJO staff proved invaluable. Early in the analysis phase, customer service paths at 4FIZJO and the most common types of inquiries and problems were meticulously mapped out. This allowed for the precise tuning of the AI's operation to the company's realities. The owners of 4FIZJO emphasize that collaborating with experienced specialists is crucial to avoid mistakes and accelerate development—using proven solutions and experts from the outset opens up new possibilities for the company. In this project, the close contact between 4FIZJO and AskSpot (frequent workshops, feedback for the AI models, iterative improvements) ensured that the agent was "taught" the specifics of the industry and the company. This approach resulted in above-average implementation results—the AI not only relieved the team but actually improved the customer experience by providing fast and accurate answers while maintaining a human context and empathy in communication.

‍