Rabona AI, a Tokyo-based company led by CEO Sho Toribe, has launched what it describes as Japan's first blended AI call center. The service uses a single AI agent to switch between inbound and outbound calls depending on the situation. It combines natural speech technology from the University of Tokyo with automated post-call transcription, summarization, SMS delivery, and system records.
A Response to Fragmented AI Phone Systems
Many companies still face workflow interruptions and callback backlogs even after adopting AI phone tools. Traditional systems are often separated into inbound-only and outbound-only functions, so follow-up calls and reminders still require human staff. Rabona AI aims to move beyond scripted answers by allowing the AI to confirm customer data, accept requests, and update records during the call.
The company has built this model on experience from more than 50 client companies and over 500 million calls. Its previous services include AI telemarketing, immediate real estate response calls, and delivery date confirmation calls. That foundation informed the development of an AI agent that completes operational tasks rather than only answering questions.
Core Capabilities of the Blended AI Call Center
The blended design prioritizes inbound calls during peak periods and shifts to outbound tasks such as callbacks, reservation confirmations, and reminders when volume stabilizes. Because the AI carries information from incoming conversations into outgoing calls, customers are not asked to repeat the same details. The system can also handle multiple simultaneous calls, reducing missed calls during busy periods.
When a call arrives, the platform identifies the caller through CRM, SFA, or core system records and retrieves the person's name, contract details, and previous inquiry history. Outbound calls are tailored to each customer based on that information, enabling personalized conversation. Unlike IVR systems, the AI does not ask callers to press numbers, and it supports SMS-based identity verification where needed.
Integration with kintone, HubSpot, Salesforce, Slack, LINE WORKS, Microsoft Teams, and Chatwork allows call summaries and outcomes to be recorded automatically. The AI handles customer data lookups, delivery date guidance, reservation intake, registration, and SMS or email sending. Calls are recorded, transcribed, and summarized, with urgent matters transferred to human staff.
Customer Results and Pricing Model
Homiya Setsubi, an air conditioning and water heater installation company in Kanagawa Prefecture, uses the service for both post-installation follow-up and installation date confirmation calls. The company reports that the AI completes around 90% of inbound follow-up calls without human intervention. It has also automated outbound confirmation calls, including leaving voicemail messages, which saved hundreds of hours of staff time.
ZAP Inc., a light cargo delivery and third-party logistics provider in Osaka, uses the AI to confirm delivery dates and manage callback calls from customers who missed initial contact. The company reports an approximately 80% reduction in phone handling related to delivery dates. Drivers can now focus more on deliveries, and the operation has expanded from Osaka to Saitama.
Rabona AI is also introducing an outcome-based pricing plan for some customers. Under this model, charges reflect completed results such as confirmed delivery dates, resolved inbound inquiries, or booked appointments, rather than only user counts or call minutes. The company will continue offering its standard monthly base fee plus usage-based pricing.
By combining inbound and outbound call handling with customer data integration and natural speech, Rabona AI is positioning its AI call center as an operational agent rather than a simple answering tool. Early deployments suggest that the service can reduce routine phone workloads and allow employees to focus on complex or judgment-based tasks. The launch reflects a broader shift toward measuring AI contact center value by completed outcomes and resolved customer requests.