Logo UEHP

European Union of Private Hospitals

EPHA Awards 2023 – Innovation in Health

1st prize, Lusíadas Saúde (Portugal), for LUSI – Lusíadas Digital Assistant

The project aimed to implement a Digital Assistant to answer customer requests supported in an omni channel platform: answering customer requests by voice (Contact Center) or text (in WhatsAPP and Facebook Messenger).

The objective was to provide a superior customer experience in Customer Service, allowing a self-service solution available whenever the customers wanted and throughout the channel customers felt more comfortable to use. In a nutshell: Increase availability (24×7), convenience (whatever way the customer wants) and with a superior customer experience (an automated self-service but human-like experience that is effective and accessible).

Using AgentifAI’s solution (our partner) we co-create “LUSI” – Lusiadas’s Digital Assistant, which uses cutting edge advanced AI technology with Machine Learning processes (ML), Natural Language Processing (NLP) and is supported in a cloud based infrastructure with an open architecture, that allows a flexible, scalable and fully available solution.

In an implementation project of ~6 months, we deployed “LUSI” in the Contact Center. The bulk of the initial work was a backoffice effort in data structuring, system integrations and APi deployment to be able to interface with the “new front-end” (Voice and Text Assistant) with our core services; and them we deployed LUSI (the AI Assistant), teaching it to understand the requests from customers and the answers to be given and to execute their requests, all in a fluent an human-like conversation.

The deployment was done sequentially to all Lusíadas Units, and today LUSI is answering all voice Contact Center calls and Facebook Messenger and WhatsAPP chat queries from customers.
LUSI was the 1st Digital Assistant in the world that was able to autonomously manage booking of exams and medical appointments, in a natural conversation with customers, as well as change, reschedule and cancel those appointments. It manages the complexity of doctor’s agenda availability, historical background of patient attendances, availability of medical specialities in several units in different locations, etc. All the complexity is handle by LUSI and allows a smooth and natural interaction with customers.

Beside all of the above, LUSI is also capable of answering all sorts of customer queries: clinic locations, addresses, payment information, timetable availability, medical specialities available in a specific healthcare unit, etc. It gives answers to most of the customer’s requests and, if it’s not able to answer (or for business reasons, is programmed for it), LUSI escalates the conversation to the adequate human assistant to handle it.

Currently LUSI is capable of understanding more than 88% of customers requests and answering autonomously more than 50% of their requests; The majority of requests that are not answered automatically are either because it is something that we defined that we wanted to escalate or because we do not have API services in place to enable LUSI to execute the request of customer (e.g. create a new customer in database). In these cases the request is handed to the right human assistant that can complete the service requests, maintaining the full context of the previous conversation with LUSI, therefore enabling a more efficient and smooth experience to customers.

In December of 2022, LUSI was extended to assure 24x7x365 customer service (even out of Contact Center working hours) and allowing a fully convenient Customer Service; till then, the service was limited to opening hours (10 hours per day, 6 days per week – with the exception of holidays). The impact after 1 year of implementation is: LUSI answered over 1M conversations, nowadays>50%of Customer Service contacts; The FTE equivalent of the contacts answered by LUSI as 48% of the Contact Center working force;LUSI was responsible for handling the booking of 45% of the digital exams and medical appointments; 100% of customer contacts out of working hours were done by LUSI.

Outcomes

This solution was aimed at improving the relationship with customers and have an impact in different aspects of healthcare access. In that context, yes, we believe the solution demonstrated significant progress, especially considering it is the first solution of AI Assistants in healthcare that provides this level of functionality it does. RPA, Digital assistants are spread but with a strong lack of reality and similitude to human behavior.

The actual impact was most visible in 5 key areas:

  • Reduction and optimization in health backlog: after covid19 screening and medical care was waiting and waiting. A serious delay in the access to healthcare comes from a variety of aspects, nevertheless access and booking was shown to be a crucial one.
  • Improved customer experience: The digital assistant system provides customers with an efficient and convenient way to schedule appointments, access information, and manage their medical records, with no waiting time and available 24×7. This results in an overall improvement in customer satisfaction and loyalty.
  • Increased efficiency and productivity: the digital assistant system significantly reduces the workload of healthcare professionals and Contact Center staff. This resulted in an increase in efficiency and productivity, allowing healthcare professionals to focus more on patient care than administrative tasks. (each operator goes from 80 calls to answer near 170 on average).
  • Reduced no-show: By automating the scheduling of appointments, proactive reminders and scheduling confirmation (with outbound calls to customers that can also automatically reschedule appointments), the solution has effectively reduced the number of patient no-shows. This resulted in a more efficient schedule that improved allocation of hospital resources making the whole healthcare system more efficient. (2 to 3% reduction depending on the unit).
  • Enhanced accuracy and reliability: The use of natural language processing and machine learning algorithms helped improve the accuracy and reliability of the digital assistant system. This led to fewer errors and a more seamless customer experience. (only 32 mistakes reported in one year).

Cost savings: a digital assistant system resulted in cost savings for the Healthcare Group that we can estimate that can reach 1M€/per year only in direct costs.

Positive sides effects

With the automatization of conversations with patients, allowing a 24×7 customer support and deploying more convenient, simple and easy to use channels (WhatsAPP, Facebook Messaging), the impact of increased availability to support patients, is proving an enduring and structural impact in customer service productivity (with an automation rate that is growing the deploy of new use cases and a better accuracy), cost reduction (less costly resources) and increase revenue (with better optimization of the medical resources), as well as a reduction on waitting time in physical units.

We calculate a reduction of 15% of car/metro/other transportation to go to a unit to book appointments.

The sustainability of the initiative is visible. e.g., in the number of tickets (conversations managed by LUSI) that has grown +47% in 1 year or the number of bookings handled by LUSI that has grown +48% in the same period.

Additionally the technical capabilities for managing the Digital Assistant is also improving as new use cases are added continuously, increasing the capabilities to answer autonomously new requests with new flows and with clinical pathways for the patients to not go to the hospital, mainly in chronic pathologies, if it is not needed.

https://www.lusiadas.pt/lusi-assistente-digital-lusiadas