For AIBUS, AI only truly matters when it solves specific problems in business operations. From that perspective, AIBUS began the development journey of AZbot to address sales challenges, initially by helping businesses manage customer conversations. Real-world deployment later showed that answering messages was only a small part of the sales process. The greater value lies in enabling AI to participate more deeply throughout the entire consultation and sales journey.

AIBUS started with business problems, not technology

Anyone who has sold products on Facebook, especially as customer volume begins to grow, will likely encounter several familiar problems:

• Messages pile up during peak hours and cannot be handled quickly enough.
• Customers have to wait too long for a response.
• Some conversations are missed when inbox volume becomes too high.
• Sales teams spend significant time answering repetitive questions.

These are problems where AI can play a meaningful role.

AIBUS saw the potential to use AI to help online sellers respond faster, make better use of product data, and reduce repetitive work for sales teams. AZbot was developed from this approach.

However, the goal from the beginning was not simply to build another chatbot that could send automated replies. AIBUS wanted to explore a broader question: how deeply can AI participate in a real-world sales process?

Real-world deployment helped AIBUS discover new challenges

Starting in the second quarter of 2025, AIBUS began researching and developing AZbot as an AI solution for online sellers. Drawing on the capabilities and product development experience accumulated across multiple projects, AIBUS focused directly on one of the most difficult challenges: enabling AI to truly understand and support a real sales process.

An AI product may perform well in a testing environment but encounter many different situations once it starts interacting with real customers.

Some customers ask directly for the price. Others describe their needs in detail. Some begin by asking about one product and then switch to another. Some only need a simple piece of information, while others require multiple rounds of consultation before making a decision.

Each industry also has its own sales process.

Cosmetics stores need to understand skincare needs. Fashion stores need to handle questions about sizes, colors, and inventory. Accessories stores need to identify the product or device a customer is currently using before recommending compatible options.

Deploying AZbot helped AIBUS realize that an AI sales system cannot rely only on a good set of answers. It needs to understand each business’s data, recognize the context of a conversation, and apply the appropriate consultation approach for each situation.

The challenges that emerged in real-world use became an important source of insight for AIBUS to continuously refine both the technology and the product design.

Developing AZbot from a chatbot toward an AI Sales Employee

If the goal is only basic customer service, a system can focus on retrieving the right information and answering questions. Sales, however, is much more complex.

When a customer is interested in a product, a salesperson often needs to ask additional questions about their needs, budget, current products, or desired outcomes before making an appropriate recommendation. If the customer does not purchase immediately, the sales process may still continue. The salesperson may need to follow up and return to the conversation at the right time.

Through these real-world situations, AIBUS recognized the limitations of chatbot models focused primarily on responding to messages. Instead of asking only, “How should AI respond to the customer?”, the team began focusing more on “What does the customer need, and what should the next step in the sales process be?”

From this perspective, AI needs to be capable of receiving customers, identifying their needs, using business data to provide consultation, recommending suitable products, following up with customers who have not yet made a decision, and knowing when a situation should be handed over to a human employee.

This became the foundation for AIBUS to develop AZbot toward the model of an AI Sales Employee.

AZbot became a platform for AIBUS to validate AI in real business environments

The AZbot development journey has given AIBUS more than just a product.

Every real customer conversation provides an opportunity for the team to better understand how people interact with AI, which situations the technology handles well, and which limitations still need improvement.

Through this process, AIBUS has accumulated experience in important areas of real-world AI product development, including organizing business data, controlling response quality, designing interaction flows, coordinating between AI and human employees, and operating systems at higher interaction volumes.

This represents a major difference between building an AI demo and making AI a practical part of a company’s daily operations.

Capabilities AIBUS has built through the AZbot development journey

AZbot also reflects AIBUS’s approach to technology product development.

The team does not simply identify a problem and build a fixed solution. Products are introduced to the market, real user behavior is observed, weaknesses are identified, and the system continues to evolve.

This process has helped AIBUS connect three elements that are often treated separately in AI development:

• The first is technology, meaning the capabilities of AI models and the supporting systems behind them.
• The second is product, meaning how technological capabilities are transformed into an experience that users can work with every day.
• The third is business, meaning the final product must create a meaningful impact on business operations.

AZbot is one of the products through which AIBUS continuously validates the connection between these three elements.

The AZbot development journey and AIBUS’s adoption of AI are still continuing

From an initial challenge of managing sales conversations, AZbot is gradually evolving toward an AI Sales Employee capable of participating more deeply in business processes.

For AIBUS, however, the significance of this journey goes beyond AZbot itself.

The experience gained from building and operating the product provides a foundation for AIBUS to continue exploring how AI can participate in other areas of business, from sales and marketing to customer service and operational automation.

AZbot is therefore both a product and the result of an approach AIBUS continues to pursue: start with a real business problem, bring technology into actual operations, learn from the market, and keep improving until AI creates clear and measurable value for the business.