Ronlight is the sole distributor of Garmin products throughout Israel, as well as several other popular consumer electronics brands. Ronlight made the decision to implement Over.ai's AI-powered representatives into their contact center in order to raise efficiency, reduce costs, and improve customer satisfaction. Avishay Pariz, CEO of Garmin IL, discusses his thoughts and experiences and shares his advice to other business leaders in the age of automation.
Q: What led to your decision to automate your contact centers with a Voice AI-powered agent?
Obviously a large part of the decision was the cost saving potential, but our main goal was to improve and grow our customer service. We knew that in order to scale our business, we had to find a way to scale our customer support in order to meet the needs of a growing customer base.
We noticed that a lot of the customer inquiries and complaints were very similar and repetitive, and our customer support team was providing the same answers over and over again. We recognized that if we could just create a template to address these repetitive issues and automate this process in some way, we could probably save a lot of customer service costs and hopefully also improve the level of service we give to our customers.
Personally, I’m a big fan of virtual assistants, I have many IoT products at home and I’m a part of the voice technology community in Israel, so that’s how I heard about Over.ai. I saw the growing capabilities of voice assistants and have seen what roles they are filling globally. There is a lot that voice assistants can do, I think that one of the first and simplest ways that voice assistants can be implemented is to improve customer experience.
Q: Which call processes did you choose to automate and for what reason?
My first instinct was — I want to automate everything.
Obviously we needed to be realistic and prioritize the implementation, so we decided to start with the things that are most commonly asked before moving on to more complex questions. We started by automating answers to frequently asked questions, such as "Where is the nearest point of service?", "What are your opening hours?", "How can I download the user manual?", etc. — regularly asked questions that have an exact, specific answer.
Q: What was the implementation process and pilot phase like?
It was above all expectations. I have experience with these types of projects, and I expected that a "three month project" would take much longer. The process was much faster than that and we went live after just a short preparation period. At first, we were in "stealth mode", which means the bot answered 10% of calls. We listened, we improved it, and every week we increased the number of calls. Less than 4 months from our first meeting, we were live with 100% of the calls.
After perfecting the most basic automations, we began adding in more complex answers every week or two, and now the system is fully integrated with our CRM and ERP. This allows people to check the real-time status of their Garmin watch in the service center, get guidance on how to fix things, and even receive manuals and links to specific videos via SMS.
Now that our employees have more time to focus on other tasks, we were able to launch a youtube channel, upload all of our manuals to the web and allow Roni, our virtual assistant, to send callers an SMS with the exact information they need.
Q: What do your results look like at this point in time?
In the first 3-5 months, we managed to implement automations to answer around 70% of caller inquiries.
In the dashboard that Over.ai built for us, we see the daily live numbers — the results are amazing. On average, Roni successfully answers more than 80% of the calls with absolutely no human input. Duration of calls dropped from around 2.5 minutes to less than 1 minute, we have 0 waiting time and soon we are going to use this success to actually promote our business with a marketing campaign that our call center is now open 24/7.
Before, we were only open 9 hours a day, 5 days a week, with a waiting time of 1-2 minutes before someone answered, and the average call duration was around 3 minutes. All of our numbers have improved dramatically, customers are receiving quick resolutions and we are testing it with digital surveys to our customers — we’re getting very good feedback regarding Roni.
Q: How do you measure success of calls?
At the end of every call, Roni asks the customer “Is there anything else I can help you with?” to confirm that the customer feels their issue has been fully resolved. On the chance they say yes, then the caller will be transferred to a human representative.
If it is in the off-hours, Roni will take the lead and a human representative will get back to them the next business day. Only around 20-25% of calls are transferred to a human representative.
Q: How has your Voice AI agent impacted average call duration?
Call duration has dropped significantly, meaning that our customers are receiving the information they need faster and are getting efficient resolutions to their inquiries. Before implementation, we had an average call duration of around 2.5 minutes; now, we're down to an average of under 1 minute.
Q: How have virtual agents impacted call hold time?
This is one of the best parts — hold time has dropped to 0, and every single customer is answered immediately.
Humans have limitations — they can only answer a certain number of calls a day, they need to take breaks, eat, go to the bathroom, etc. — but Roni doesn’t have limitations.
Q: What’s been the impact of your Voice AI agents on your human contact center employees?
We didn’t fire anyone — we switched people to other positions in the company. It took a little while to get used to, but our employees have learned to work alongside Roni, and have understood that the automation of those processes makes their jobs better, more efficient, and truly allows them to spend their time on things that make a bigger impact.
Q: How do your customers react to your Voice AI agent?
Before implementation, people were calling the call center multiple times — waiting on hold for a few minutes, hanging up, then calling back later. Now, we’re seeing this isn’t happening as much, because people understand that they can get their issue resolved the first time — it's one touch customer service.
It’s very hard to measure feeling. We listen to calls, make sure that customers are happy and not angry and we send them surveys in SMS to get additional feedback.
For some customers it takes time to adjust and a few of them have asked to speak with a human representative, but after a few months both our customers and our team are learning how to interact with Roni to get quicker resolutions for everyone.
Q: What does the future look like for Garmin Israel and Over.ai?
Our wish is to deepen the integration, give Roni more insights and more robust knowledge to be able to answer even more complex questions. Right now, she gives somewhat general answers — in the future I want her to be able to give more technical, in-depth answers.
It takes time to train a human to answer questions in such a technical, knowledgeable manner, so we are patient and we are looking forward to teaching Roni a lot more.