Perxona AI is an enterprise product and does not provide a standard process through which ordinary users can register and immediately test the complete experience.
More accurately, the observations here are based mainly on public demos and deployed use cases.
I was particularly interested in “JoJo” from Japan Boost Travel.
Rather than appearing as a small chat bubble hidden in the lower-right corner, it interacts with travelers through a visible character. Users can ask questions such as “Where should I go on my first trip to Japan?” or “What do you recommend nearby?” The character then continues answering based on the platform’s content.
Compared with an ordinary chatbot, its visual presence is significantly stronger.
The character nods and uses facial expressions and gestures. These movements are not particularly complex in isolation, but within a conversation, they make the interaction feel more like being “received by someone” than using a text box.
This is also the most noticeable expression of Behavior AI™.
However, that proactive quality is a double-edged sword.
Travel websites naturally involve questions and recommendations, so a digital human proactively asking about a visitor’s needs does not feel out of place. On a product page where the user only wants to check a price quickly, however, a character that repeatedly appears and speaks may become a distraction.
The usefulness of Perxona AI therefore depends heavily on when the character appears.
A more proactive digital human is not automatically a better one.
The second issue is the knowledge base.
No matter how natural the Avatar appears, users will not forgive incorrect product information, outdated prices, or confused recommendation logic simply because it can smile.
In a real implementation, knowledge content and back-end business connections may be even more important than the character’s appearance.
Pricing is another concern.
Perxona AI does not publish prices. Given the nature of the product, the cost is unlikely to cover only a software account. Brand-character development, the knowledge base, system integrations, and ongoing maintenance may all become part of the project cost.
This may be acceptable for large travel, retail, or exhibition projects, but small teams cannot easily purchase a one-month trial with a credit card as they might with an ordinary chatbot.
Pros
- 3D interaction has a stronger presence: Compared with a standard text-chat interface, character expressions and movements are better suited to customer-facing scenarios.
- Can provide proactive service: It does not need to wait for users to ask questions and can offer assistance at appropriate moments.
- Suitable for brand characters: A mascot can move from being a static image to a digital character that actively participates in service delivery.
- Well suited to multilingual scenarios: Tourism, hospitality, and cross-border retail offer particularly clear use cases.
- Does not depend on traditional motion capture: Behavior AI™ can automatically generate some actions and expressions in response to the conversation.
Cons
- Pricing is not transparent: You must contact sales to learn the actual project cost.
- Deployment is more involved than with a standard chatbot: The Avatar, knowledge base, and business systems all need to be considered together.
- Excessive proactivity may disrupt users: Trigger logic and interaction design are important.
- Results depend heavily on the knowledge base: A natural-looking character cannot compensate for incorrect or outdated information.
- Public use cases remain concentrated: Current examples are more commonly found in tourism and Asian markets.
Best for / Not ideal for
Best for
- Tourism and hospitality: Digital humans are well suited to handling frequent questions, recommendations, and multilingual reception.
- Exhibitions and physical displays: A digital character can attract attention while also answering questions.
- Retail and e-commerce brands: Suitable for businesses with a clear brand IP that want to turn the character into a shopping-assistance interface.
- Corporate training: A consistent character can introduce courses, answer knowledge questions, and provide internal support.
- Medium-sized and large organizations with clearly defined service scenarios: They can integrate the Avatar into existing business processes rather than using it only as a demonstration.
Not ideal for
- Businesses that only need a basic Q&A bot: If text chat is sufficient, a 3D Avatar adds unnecessary cost.
- Small businesses with limited budgets: Custom pricing and project-based deployment create a relatively high entry barrier.
- Businesses with very simple webpages or customer journeys: Proactive intervention by a digital human may create more friction than value.
- High-risk scenarios requiring strict control over AI output: Proactive recommendations and LLM conversations require a more comprehensive review framework.
- People looking for an immediate self-service trial: This is not an ordinary SaaS product that can be fully deployed within minutes of registration.
Comments (0)