After moving to Spain, I spent too much time on calls about luggage, visas, and getting internet set up. The task was often simple. Getting it done in another language was harder.
When I heard about AI assistants making calls for people, I wanted one. I also wanted to hear the conversation, follow what was being said, and guide the assistant when a new detail changed the plan. That became the starting point for Bombyx Calls.
Feedback on the calling system brought healthcare language access into focus. We are now discussing integrations to handle routine patient calls across languages, with staff guiding the conversation. The purpose is clear: help patients reach the right support and reduce repetitive work for staff.
Stay involved while the assistant speaks
AI calling is becoming part of everyday services. In July 2025, Google announced that Search could call local businesses to ask about prices and availability. Collecting that information saves people a round of phone calls.
I wanted help with the decisions that come up during a call, too. If a bicycle workshop offers an appointment but needs to keep the bike overnight, I need to decide whether that works. Instructions given before the call cannot cover every turn in the conversation.
With Guided AI Calls, the assistant speaks while you listen and follow the transcript. You give it private guidance as the call unfolds. On supported Spanish calls, you can read an English translation alongside the original transcript and give your instructions in English.
You stay responsible for the choices. Available destinations and languages appear before the call starts, and you should check important details.
Put the same control in your team’s hands
A business faces the problem from the other direction. Someone calls with a question, but the person who knows the answer is busy. Staff spend time gathering details, repeating basic information, and finding the right colleague.
Incoming Calls handles routine questions using information the business approves. Staff follow calls, give private guidance when needed, and handle decisions and follow-up. We set the service up around the team’s work.
For multilingual services, that setup includes the caller’s language, the staff’s working language, and the translation between them. We agree and test the language coverage for each service. The aim is to let the same team help more callers without every routine question becoming a transfer.
Help patients get through in their language
A patient needs to change an appointment but does not speak the receptionist’s language. They need to explain the request and understand the next step. The booking team needs enough information to act without making the patient repeat the story.
Health systems already use telephone and video interpreting to connect patients and staff. An interpreter carries the meaning between languages. A clinician makes care decisions.
AI is entering the routine part of that work. LanguageLine describes a pilot for low-risk AI interpreting with access to human interpreters. Its service shows how routine exchanges and professional interpretation can sit within the same call path.
Our focus is the whole journey: answer the routine question, involve the right staff member, and leave the patient with a clear next step. Private guidance keeps the team’s knowledge and decisions connected to the conversation.
A patient speaks Arabic. The booking team works in English.
Our healthcare calling page lays out the service we design around a health system. Here is the appointment-change example from that plan.
The assistant confirms Arabic as the patient’s preferred language and gathers the request. A booking-team member reads it in English and privately directs the assistant to ask whether Friday morning works. The assistant asks in Arabic and saves the request for the team to check.
The booking decision stays with the team. The patient hears that availability will be checked, rather than being promised an appointment that has not been confirmed.
Care questions need their own route to a clinician and qualified interpreter. A staff member reading an English translation cannot confirm that every Arabic word was translated correctly. Clinical judgment and language quality each need the right people.
What is built, and what needs testing
Guided AI Calls has a personal-call flow and a scripted practice experience. Incoming Calls has a staff call desk, a fictional browser simulation, and a telephone demo. You can explore both calling directions before planning a service.
The healthcare journey describes a tailored service. Arabic in the example is a language to agree and test with the health system, not a claim that a finished Arabic healthcare integration is running. Setup covers approved answers, staff roles, interpreter handoffs, patient information, and the systems the team uses.
We have not measured patient outcomes, staff savings, or shorter waits from a Bombyx healthcare deployment. Those results need evidence from the service in use.
Measure whether callers get the help they need
For patients, success means getting useful help in their language. For staff, it means less repeated work and enough context to handle the next step. A natural-sounding voice alone proves neither.
The measures are concrete: time to useful help, time to the right person, completed requests, repeat calls, and language quality. Staff review, interpreting, follow-up, phone service, and AI all count toward the cost. A fast greeting matters less than a request completed correctly.
This is the kind of system we build at Bombyx Labs: an AI worker connected to the people and information needed to finish a real task. In this case, a personal frustration led to a calling tool, and feedback led us to a wider business need. The same principle runs through both: the assistant handles the conversation while people keep control of the decisions.
You can explore Bombyx Calls and try the practice experiences in your browser. If your team spends too much time repeating answers or moving calls between people, bring us that call workflow. We will help you identify the work an assistant should handle and the decisions your team should keep.