"Aurora AI Call" Achieves Flexible Automated Response by Combining AI Voice and Button Operations; New Call Flow Feature Enables On-Site Design

Media4u Inc. (Minato-ku, Tokyo; President & CEO: Toru Iwadate; hereinafter "Media4u"), a wholly owned subsidiary of Fabrica Holdings Co., Ltd., has added a new feature to its voice AI agent "Aurora AI Call" that allows users to intuitively build call flows without writing prompts. This feature enables seamless switching between flexible "AI voice dialogue" and reliable "DTMF (push-button) branching," simultaneously improving customer experience and operational efficiency.

Recently, there has been growing expectation for AI voice response in call centers and corporate phone desks to address labor shortages. However, barriers such as "difficult setup" and "not knowing how to write AI prompts" have been major hurdles to adoption.

With this update, even those without specialized engineering knowledge or those who have struggled with tool onboarding in the past can intuitively design and build call flows that match their business needs.

Background of the Update: Solving Challenges Across Industries by Combining the Best of AI and IVR

Traditional phone automated response systems, such as "dial-push IVR" and "AI agents" that converse in natural language, each have their pros and cons. Push-type IVR can reliably branch requirements, but long voice announcements and menu hierarchies can lead to customer drop-off and dissatisfaction, making flexible responses difficult. On the other hand, AI agents can engage in flexible, human-like conversations, but for simple, routine inquiries (e.g., store hours, requesting standard documents), they can be overkill, leading to unnecessarily long call times.

This update achieves the best of both worlds. It allows companies to flexibly combine the optimal approach based on their challenges, the use case (inbound/outbound), and the complexity or urgency of the customer's inquiry.

Strengths of the Hybrid Approach

■ Ensure Required Procedures Are Followed
Steps such as identity verification, mandatory announcements, and compliance-required processes are explicitly controlled within the conversation flow, structurally guaranteeing that essential tasks are completed. This ensures safe use even in operations requiring legal compliance.

■ Conversations as Natural as a Human's
Within each step, the LLM (Large Language Model) flexibly responds by understanding the context. It handles unexpected customer questions or phrasing with natural, non-robotic dialogue.

■ Reliable Input via Push Dial
Input of numbers such as customer IDs, dates, and amounts is reliably handled via push-button (DTMF) input, not voice recognition. This eliminates concerns about mishearing in noisy environments and handles information that is difficult to speak aloud.

■ Simple Modification and Operation
Specific steps can be replaced or improved without rebuilding the entire flow. The speed of specification changes and tuning significantly reduces operational costs.

■ Traceability of Responses
Because it is clear what happened at each point in the flow, quality improvement and audit responses are easy. The system is not a black box. Talk, listen, press. All operations combine reliability and naturalness.

Industry-Specific Use Cases for Hybrid Design

By combining "AI x push-button branching," more practical and effective automation is possible across a wide range of industries and scenarios.

Financial Institutions & Insurance Companies (Customer Centers)
Urgent matters like "lost or stolen cash cards" or procedures requiring strict information management and identity verification, such as "PIN inquiries," are transferred to a dedicated desk via push-button operation. Meanwhile, "general service overviews" and "document requests" are handled flexibly by the AI agent in a conversational format, providing a zero-wait-time customer experience.

Real Estate & Property Management Companies (Resident Support Desk)
Emergencies like "water leaks or lost keys" are immediately transferred via push-button to a 24-hour emergency desk (staffed). Tasks like "moving-out procedure applications" or "parking certificate issuance requests" are handled by the AI agent for initial intake, which thoroughly collects necessary items (property name, name, desired date, etc.) and leaves a summary.

Automotive & Inspection Centers (Reservations & Inquiries)
Urgent matters like "road service requests due to accidents or breakdowns" are prioritized for connection to a specialist desk via push-button. Routine questions like "store hours and location" are answered by immediately sending an information URL via SMS. Tasks requiring detailed information, such as "reservations for vehicle inspections or regular maintenance" or "quotation consultations," are completed by the AI agent through dialogue, significantly reducing the phone burden on store staff.

Main Update Details:

■ Node-Based Visual Conversation Flow Editor: Dramatically Streamlines On-Site Onboarding
Users can intuitively design call flows by dragging and dropping nodes (processing units) like "utterance," "branch," and "action" onto a canvas on the management screen and connecting them with lines. Transition conditions between nodes can also be visually set and confirmed, making the conversation path with the customer clear at a glance. This allows for easy creation and modification of even complex response flows. This significantly reduces the time and effort required for on-site staff onboarding. Even new managers or operators without system expertise can intuitively grasp the overall flow, ensuring smooth handovers and training. Creation and modification of complex response flows due to daily business changes can be completed on-site without relying on engineers or IT departments.

■ Push-Button Branching and Action Execution: Supports Operations Requiring Strict Accuracy (Finance, Security, etc.)
When a caller operates a phone button (e.g., pressing "1"), pre-registered actions (e.g., call transfer to a representative, sending a detailed URL via SMS) are executed immediately. This enables efficiency gains, such as guiding simple routine questions to SMS via push-button while leaving complex consultations to AI. It is also highly effective in areas where "AI hallucinations (incorrect answers) are absolutely unacceptable" from a compliance and security perspective.

[Use Case Example: Hybrid Operation at Financial Institutions and Insurance Companies]
Phases requiring strict information management and 100% accurate processing, such as account inquiries, PIN entry, and contract changes, are handled reliably via "push-button input" or "transfer to a dedicated desk." Meanwhile, general service overviews and document requests are flexibly handled by the "AI agent" in a conversational format. This achieves a hybrid design that ensures security while reducing customer wait times.

What is "Aurora AI Call"?
"Aurora AI Call" is a cloud service that automates inbound calls, outbound calls, and multilingual support using an AI agent. It operates 24/7, tirelessly supporting customer phone responses.