Digital humans for customer service: when they fit and how to start

If your organization serves people face to face, at a front desk, a public service office, a showroom, an exhibition booth, or a museum, you have probably seen the same pattern: a large share of your staff's day goes to the same handful of questions. An AI digital human is one way to take on part of that load. It is not the right answer for every service point, though, and it is worth understanding where it fits before deciding.

This article is written for decision makers who are weighing that choice. We use AIVA, Digibay's AI digital human, as the reference.

The challenges at face-to-face service points

Most service points we look at share four challenges.

• Limited service hours. Service depends on staff availability at each location. When staff are busy, on a break, or the counter is understaffed, visitors wait or leave without an answer.
• Repetitive questions. Staff spend much of their time on basic questions: locations, schedules, requirements, and prices. These are important to the visitor but rarely need a person's judgment.
• Inconsistent information. Answer quality depends on who is on duty. Two visitors asking the same question on different days can leave with different answers.
• Information media that gets ignored. Brochures and static screens do little to catch visitors' attention. The information is there, but people walk past it and queue at the counter instead.

What an AI digital human does

AIVA is an AI virtual assistant in the form of a 3D character that interacts with visitors through voice conversation. A visitor asks a question out loud, and AIVA answers by voice, with facial expressions, body movement, and lip-sync that follow the answer. Unlike a text-based chatbot, it is visually present at the service location.

Behind the character, AIVA works in four steps. The visitor asks a question, the voice is converted into text, the AI finds the answer in the organization's knowledge base and related system data, and AIVA delivers the response with voice and expression. The knowledge base holds FAQs, product information, procedures, and schedules prepared by the organization.

AIVA can also connect to existing systems through APIs, so some answers follow the latest data. Examples include a visitor's current queue number from a queue system, product availability from an ERP or inventory system, a doctor's schedule from a scheduling system, or application status from a CRM or service system. If you want to understand how these layers fit together, our article on designing digital humans as system interfaces covers the technical side.

Where it fits

A digital human tends to fit when a service point has many recurring questions, a steady flow of visitors, and information that can be written down clearly. Typical settings include:

• Customer service and front desks: greeting guests, answering common questions, and directing them to the right service, in office lobbies, banks, and hospitals.
• Public services and government offices: explaining requirements, service flows, and queue status to the public.
• Showrooms and retail: helping customers find products, compare options, and leave their details as prospective buyers.
• Events and exhibitions: drawing visitors to the booth, introducing products, and collecting visitor data.
• Museums and education: presenting information about collections, history, or learning material interactively.
• Hotels and hospitality: sharing information about facilities, room services, and nearby recommendations.

AIVA has been presented at public settings such as "NYALA - 200 Tahun Perang Diponegoro" at Museum Galeri Nasional Jakarta and World AI Show 2025 at Hotel JW Marriott Jakarta.

Where it doesn't fit

Being clear about limits is part of a good decision. A digital human answers within approved information and permitted systems. It is not a replacement for your team.

• Complex or sensitive cases still go to staff. Complaints, exceptions, negotiations, and anything that needs judgment or empathy should be handled by people. A well-designed digital human recognizes when a question is outside its scope and points the visitor to a staff member.
• Information that is not documented can't be answered well. If procedures live only in people's heads, the first step is writing them down, not installing a screen.
• Very low-traffic points may not need it. If staff already have time for every visitor, the case is weaker.

The goal is to let staff spend less time on repetitive questions and more on the cases that need them, with humans in control of the service.

How answers stay controlled

For most organizations, the first concern is whether the AI will say something it shouldn't. AIVA addresses this in three ways:

• Answers are limited to the knowledge base the organization has approved.
• Access to systems is set according to permissions, so AIVA can only reach the data and functions it has been allowed to use.
• The knowledge base can be updated as needed, so when a schedule, requirement, or product detail changes, the answer changes too.

This also helps with consistency. Every visitor gets an answer from the same approved source.

How to start: implementation steps

We recommend starting with a pilot at one service point, learning from real visitor questions, and expanding from there. AIVA implementation follows four steps.

1. Discovery. Analyze needs, the location, and the service flow. Which questions come up most? Where do visitors stand? Which systems hold the data they ask about?
2. Preparation. Build the knowledge base, design the avatar, and set up the screen display. The avatar's character, clothing, and attributes can match your brand identity, and the screen can include an interactive menu, product catalog, QR code, or lead form.
3. Installation and training. Install the device, whether a touchscreen kiosk, LED screen, hologram fan, or custom booth, and train the people who will use and manage it.
4. Operations and support. Maintenance, support, and system updates, including keeping the knowledge base current.

A pilot keeps the scope manageable. It shows which questions visitors actually ask, where the knowledge base needs more detail, and how staff and the digital human work best side by side. Because AIVA is developed by Digibay, one team handles the avatar, software, system integration, and devices.

Talk to us

If you run a service point and want to see whether a digital human fits, we are happy to walk through your questions, location, and systems with you. Contact Digibay to schedule an AIVA demo.