By deploying Erica, their Observe.AI VoiceAI Agent, the digital health provider eliminates missed calls, reduces backlogs, and frees staff to focus on complex cases while upholding compliance.
Digital health provider Simple Online Healthcare needed to balance rapid increases in medicine orders with maintaining the quality of its patient interactions. In the UK, where the company is based, demand for weight care medication often peaks at the end of each month, when patients rush to place orders after their payday.
“All of a sudden, we have this backlog of orders,” says Nikita Lavrenov, Special Projects Associate at the company, which began with over-the-counter treatments before expanding into online general practitioner services.
“We need to process these, and we need to ensure that we still provide the best quality of representing our company and speaking to our customers, answering all their frequently asked questions or concerns regarding their order.”
Simple Online Healthcare is growing rapidly month over month, which has increased the need to support a larger customer base. The rise in demand from new business grew call volumes, and many prospective customers were dropping out of the IVR before reaching an agent due to hold times and limitations of the system for routine queries.
“We were operating at multiple layers,” shares Arnav Dev, Product Manager at Simple Online Healthcare. “An interaction with a patient often requires frequent follow-ups to make sure the service we’re offering them is safe. However, the healthcare industry is pretty conservative about automation, so we had to find the right entry point.”
“That made us realize the kind of quality service we wanted to deliver wasn’t reaching a big part of our patient base,” adds Dev.
These challenges created a heavy burden on the company’s contact center operations, resulting in delayed order processing and higher churn rates. To address this, Simple Online Healthcare looked to automation to increase its efficiency while maintaining the empathy and compliance required in healthcare conversations.
The company started by looking at types of calls where automation could deliver value immediately. It became immediately apparent that a bucket of routine questions covered by the existing knowledge base and ordering system would help alleviate the majority of callers’ questions, and provide a better experience without needing human intervention. . Still, many patients preferred confirming details over the phone rather than relying on self-service options.
Simple Online Healthcare rolled out a VoiceAI Agent from Observe.AI to answer all calls and triage them appropriately. “About 90% of the calls were basic questions like ‘Where’s my order?’ or ‘Can I get a status update?’” says Dev. “That’s when we realized an AI agent could take care of these while still sounding natural and empathetic.”
The AI agent, named Erica, now resolves routine questions and transfers complex or sensitive issues to human experts when needed. Unlike a simple phone menu, Erica can answer patient queries in natural conversation and draw on the company’s knowledge base to provide accurate responses. To make this work, the AI agent had to sound as approachable as the company’s best staff members. Its responses were tweaked through system prompts and informed by the company's library of patient support articles built over the years.
“The key for us was to make sure our knowledge base covers our entire Help Center, so that the AI agent only escalates queries that are beyond that quite extensive knowledge base,” explains Lavrenov.
“We were also looking for simplicity of use,” adds Dev. “Even if you’re not an AI expert, you can make updates with a simple prompt. It was one of the big reasons we went with Observe.AI, together with strong partner support and the ability to sound human, not robotic.”
The company also had to consider compliance from the very beginning, given its operations in a highly regulated healthcare environment. In the UK, General Data Protection Regulation requires call recordings to remain within the Europe, Middle East, and Africa (EMEA) region.
Observe.AI’s VoiceAI Agent supports such compliance requirements through enterprise-grade security and built-in safeguards. The solution automatically redacts personal data and enforces real-time script guidance to keep all calls compliant. Every interaction is likewise audited, with the Guardian trust framework evaluating 100% of conversations for regulatory adherence and performance monitoring.
All these controls ensure privacy is protected and that calls can be easily routed to a human agent when needed, without compromising compliance or patient trust.
Dev recalls that change management was one of the hardest parts of introducing AI into daily patient support operations. “Our company was essentially built by people coming from clinical backgrounds, not tech,” he says. “So introducing a technological shift like this meant training our people and training the AI agent to make sure we’re staying compliant and in line with our core work processes.
Patient care agents initially had to adapt to a new model in which part of their workload was now shared with the AI agent.
The company had to make sure everyone understood that the goal wasn’t replacement but to make their work easier. “It was a matter of making them realize that they were still contributing to the company. We just had to get that point across and build that trust,” says Dev.
Within months of introducing its inbound VoiceAI Agent, Simple Online Healthcare was already tracking its impact. The company monitored first-call resolution, call handling times, containment rates, and escalation levels to gain a clear picture of how the system was performing.
We can already see from data that patients are getting quicker answers and fewer transfers.”
After deployment, Erica began managing around 55% of customer calls end-to-end without escalation to a human agent.
“Prior to the introduction of VoiceAI, we had an abandonment rate of roughly 30% in IVR,” says Dev. “Since then, it’s been down to zero, because every call is being catered to.”
By taking on thousands of monthly calls and containing most routine interactions, Erica has absorbed peaks in demand, especially during end-of-month surges for weight-care medication orders, without additional staffing.
“When a new product launch triggered one of our busiest days of the year, the AI agent helped us sail through without needing to add headcount or keep patients waiting in long queues,” adds Dev.
These early results have helped the company consider how to rebalance resources between inbound and potential outbound calls, opening opportunities to reduce wait times and serve more patients.
Likewise, by taking over a steady stream of routine queries, Erica has reduced order-related backlogs and freed patient care staff to focus on cases that require empathy and clinical judgment. Patients benefit from faster answers to common questions and receive human support for complex needs.
SimpleOnline Healthcare knows that AI agents are not a one-time implementation and sees it as an ongoing effort. Data from patient interactions feeds back into the system, helping sharpen Erica’s responses and strengthen escalation rules. It also surfaces unusual cases that don’t fit its capabilities set, and flags them for further evaluation.
“We’re at the point where we’ve done enough of the inbound calls, where we started processing the data and seeing how it (the AI Agent, Erica) improves the overall call time and the percentage of issues resolved with the initial call,” says Lavrenov.
This continuous cycle of reviewing results and updating the company’s knowledge base ensures Erica keeps up with changing products, policies, and patient needs.
Dev adds, “The success of VoiceAI changed how we look at automation. Now, every new process we review starts with the question, can the AI agent take this on?”
Simple Online Healthcare is now preparing for a broader adoption of AI automation. It is extending its AI Agent, which was originally built for voice, into the chat channel. Early trials have shown that the agent can handle routine queries in chat with the same accuracy and tone patients already know from voice, while directing more complex concerns to human staff. This step allows patients to choose the channel that’s most convenient for them without compromising the consistency of their experience.
Based on this experience, the company plans to expand into proactive calls for weight care patients. Onboarding new patients in this area follows a set script and is still a manual task.
“With outbound calls, we would be alleviating work from the patient care staff by streamlining that onboarding process,” says Lavrenov.
The company is also considering expanding the AI agent's use to new regions and markets.
“We’re present in five countries and plan to expand into 15 to 20,” says Dev. “And every time we plan an expansion into a new market, VoiceAI is one of the first things we consider.”
Lavrenov adds, “Language is a major consideration, not just translation, but also the cultural differences and ways of speaking in each market. The AI agent can carry the same knowledge, but it needs the right language and dialect. With the robust language capabilities offered by Observe.AI, we’re confident to expand into new markets.”
Each new application will follow the same principle that guided the UK rollout. It will start with tasks that are high in volume and low in complexity and build trust internally, as well as with patients. From there, the company will use data to steadily broaden what Erica can do.
“The agent rollout with Observe.AI went really well and helped us reduce call handling time and resolve more patient queries on the first interaction,” says Lavrenov. “It has given us confidence to expand the solution internally and into new markets.”
.png)
.png)
.jpg)
.jpg)
.png)
