+34% Growth of Visits

Medical Voicebot

Medica Voicebot helps patients who want to quickly reschedule or simply cancel an appointment.

About client

Medica Voicebot helps patients who want to quickly reschedule or simply cancel an appointment. Availability and quick handling of the process allows for efficient updating of the facility's calendar and limit the number of visits for which patients do not show up.

Business challenge

The most popular form of contact with the doctor’s office, hospital and clinic is by telephone contact. However, it is often unreliable. A large number of callers coupled with staff shortages mean that many calls are missed, and the calls themselves are conducted in a hurry, between subsequent tasks and patient visits in the facility. This situation also affects the process of confirming appointments by patients. The staff of a medical facility is often unable to effectively confirm the patient’s arrival by telephone. This leads to a situation where many appointments are booked in the schedule and patients do not show up to the facility, thus blocking the space for other patients.

Team composition

  • project manager
  • business development manager
  • developers
  • quality assurance specialists

Our solution

The development process and our solutions

At INTechHouse, we have created a tool for automating processes related to patient service in a medical facility based on artificial intelligence technology. As part of the service, we provide a graphic wizard tool available from any browser. The user using the wizard can create and edit all processes related to the planned scenarios of conversations to be carried out by EasyBot. The wizard is divided into sections that will facilitate the stage of configuring intentions and individual stages of the conversation.

  • Flows – they are responsible for the main topics of conversation. Their configuration decides where the interlocutor will go after recognizing the assigned intention.
  • Pages – components responsible for direct communication with the interlocutor. They contain the texts of the bot’s statements, parameters and intentions responsible for understanding the context of the conversation.
  • Intentions – a key element of the structure, responsible for catching key phrases from the interlocutor’s statements and matching them to the appropriate flow or page. There are dictionaries of key phrases configured in them, which can be expanded on an ongoing basis, and thus train EasyBot so that it responds better to the diversified answers of interlocutors.
  • Entities – intent boosters. Their configuration allows you to specify the context or collect detailed data required, for example, to place an order or arrange a visit.
  • Track groups – enable configuration of groups for tracks with similar parameters, which can be a great help for complex scenarios.

Technology used in project

 

The voice assistant is built of several layers, the cooperation of which makes it able to understand speech, respond adequately to the situation and learn new things.

  • Voice interface, in particular, ASR (Automatic Speech Recognition) – it automatically recognizes human speech and the TTS (Text To Speech) synthesizer allows you to convert text back to voice.
  • NLU (Natural Language Understanding) layer.
  • Conversation layer – here is the entire knowledge base and some repetitive conversation scenarios.
  • Business logic layer – it is a kind of decision path determining what actions the voice assistant should take as part of a given instruction from the user.
  • Integration layers with external systems – connection with systems, applications, telephony and everything else that will allow the implementation of a given task.

Thanks to the mentioned ASR and STT systems, it is possible to translate speech into text, which is then transferred to the natural speech processing engine. There, a response is created, which is voiced with the help of a synthesizer or collected voice-over recordings. The whole thing would not work without an extensive knowledge base, a semantic search engine that gives you the ability to embed statements in context, and the aforementioned integrations – e.g. with the CRM/HIS system, from where the voice bot obtains the data necessary for the service process.

How about repeating the success of our clients?

Contact with us!

Value we added

Medical Voicebot, or as we call it – EasyBot – improves communication and speeds up the patient service process. The service has a built-in voice interface, thanks to which it answers the questions asked by the patient, verifies the information following the agreed scenario and helps in solving repetitive tasks.

EasyBot listens to users’ statements and, based on key phrases, matches them with the intentions previously created in conversation scenarios. When the patient calls the facility’s hotline, Easybot will collect guidelines from the patient and provide available dates, and after accepting, it will automatically register the patient in the calendar and confirm all the collected information.

Easybot will also carry out an outbound campaign to all patients booked for visits in a given facility. He will provide all the details of the booking and ask if the patient confirms arrival at the facility. The process can be carried out in several steps and combined with the sending of SMS messages in the 2Way version, i.e. with the option of writing back by the recipient YES/NO. Thanks to the integration with the institution’s HIS class system, the effects of the campaign can be automatically recorded in the calendar. Voicebot can also collect all the necessary information required to order a prescription and submit it for approval by the attending physician.

Results after starting the project 

After launching the EasyBot service on the hotline of the medical facility, the number of missed calls fell to zero. Voicebot works around the clock, receives all calls from patients and performs the processes provided in the conversation scenario for them. EasyBot increased the number of visits arranged by the hotline by 34%, which translates directly into the financial results of the medical facility. 

Another success of the EasyBot implementation worth emphasizing is the increase in the effectiveness of confirming visits scheduled in the calendar. AI Voicebot makes calls to hundreds of patients every day and confirms their arrival for a visit. Campaign statistics show that thanks to EasyBot, 90% of visits are confirmed, and the number of patients not showing up for scheduled appointments has been reduced to 5% of all visits. 

Future perspective