How AI revolutionizes the customer service team's work in telemedicine

Facing a surge in demand for telemedicine, TeleClinic found an automation solution that provided empathetic and efficient customer support.

Highlights

• TeleClinic is a telemedicine platform that revolutionizes modern medicine with 24/7 support.
• The Customer Service Team grew by 150% during the first lockdown, and will add 14 more agents in the following 6 months.
• ultimate.ai provided a flexible, customizable solution to help streamline the support request process thanks to automation and AI.
• Within 2 months of automation, 30% of support conversations were handled by the bot.

Intro

TeleClinic is a German market leader in telemedicine that launched in 2016: its platform matches patients and doctors for online consultations via a mobile app.

TeleClinic operates within a regulated and changing landscape: in Germany, telemedicine used to be covered by private insurance only.

Since the law extended coverage to everyone, there has been a dramatic surge in demand: nearly +500% in 2020.

In the same year, the pandemic also induced a steep spike in customer service requests, making it even more crucial for customer service to assist patients and doctors quickly and reliably.

The collision of these two circumstances meant it was the right time for TeleClinic to adopt automation. TeleClinic already offered 24/7 customer service, however entirely handled by live agents.

While the team has been growing fast, it became really clear that they needed a tool to address customer queries in an empathetic and efficient way.

That’s how their operations leadership decided to partner up with ultimate.ai to build it.

Challenge: Addressing a mounting number of queries in an efficient and flexible way

TeleClinic’s Customer Service deals with two kinds of requests: some are purely administrative and others are medical.

They do not have the same level of urgency, but TeleClinic’s communication system wasn’t able to prioritize the more urgent queries.

They needed a tool to do that, while ensuring a high-level customer service experience for everyone. They faced a series of challenges:

• Telemedicine is a complex, sensitive matter, where a one-size-fits-all approach doesn’t work;
• They had little actionable data on what customers needed;
• Their Customer Service team grew very fast: the first level support needed to scale up from 5 to 10-15 agents (their team is in early 2021 composed of 40 agents).
• They needed empathetic and accurate suggested replies to support their work.
• Knowing when to escalate critical queries to a human was also vital.

Solution: A made-to-measure bot to support the customer service team

TeleClinic chose ultimate.ai’s solution, because it allowed them to build and customize their bot to fit their specific needs, as well as train it themselves.

The tool is implemented into Zendesk, the CRM TeleClinic uses.

The first step was to deploy suggested replies, to help agents answer repetitive questions and train the bot at the same time.

Philipp Weidenbach was also dedicated to training the bot himself; he was guided by ultimate.ai’s customer success team throughout the whole automation journey, from initial setup to conversation design.

A month after suggested replies were activated, customer-facing automation was launched; now, the bot can deal with the most frequent administrative queries with no human intervention.

Some administrative requests (for instance to postpone an appointment) are transferred to an email ticket: the agents get notified and a 24-hour reply rule is created automatically on Zendesk.

This allows us to shift volumes, save escalation for urgent queries and help agents prioritize. The bot is also instrumental in identifying urgent requests: for instance, is a last-minute appointment an emergency?

Once the request is escalated to an agent, the bot helps them assist with more complicated questions.

And while doctors also use the chat, it was decided to always direct them to a human agent, who is guided by suggested replies, so they immediately get the assistance that they need.

Results

The automation level on TeleClinic’s chat support went from 25% in October to 37% in February.

This impressive number for a naturally complex industry was achieved thanks to the mix of suggestion and automation.

This approach continuously improves the accuracy of the bot, and ultimately, its readiness to understand and resolve incoming requests.

TeleClinic’s purpose was never to automate support entirely, but to find the right partner to create a great customer experience, especially when many end users are facing health challenges at the same time, and need empathetic and reliable help.