By Nadia Cameron - Editor - Marketing | Associate Publisher | Mi3
Adopting an Australian-made voice-to-AI assistant technology has already paid for itself – and 18 months of slogging through APIs connecting a disparate dealer management system network – by giving Suzuki Australia a way of measuring and understanding the volume and success of service calls to its national dealer network for the first time. It’s a customer retention and experience win the automotive company’s GM claims as a first for the local automotive industry and also an AI-first globally that’s now being eagerly watched by HQ in Japan.
What you need to know:
Suzuki Australia’s 95 strong dealer network has connected the service dots digitally by launching a unified voice-to-AI assistant called Hey Suzi nationally in December.
Using Australian created technology from Contact Harald, the voice and web assistance is a first for the automotive company globally.
Described as “absolute service” by Suzuki Australia’s GM, Hey Suzi addresses the biggest bugbear in the customer experience: Being able to book a service for your vehicle seamlessly, digitally and 24/7.
Among the biggest hurdles to getting Hey Suzi up and running was plugging in nine disparate dealer management systems – and one physical book system – to ensure customer and service capacity data was seamlessly accessible.
With data hygiene critical to any successful AI use case, Suzuki did a thorough review and training program for dealers to ensure data collection was as accurate and comprehensive as possible. Per GM, Michael Pachota: “Our data and data retention in terms of customer information went from 5560 per cent of customer information that wasn’t bogus, to upwards of 95 per cent in the last 5-6 years.”
Test drives, vehicle recalls, service and sales campaigns are all now further use cases for Suzuki’s AI assistant.
It’s by gaining accurate data on service inquiries and bookings across Suzuki Australia’s 95-strong dealer network that general manager, Michael Pachota, believes the automotive brand is finally tackling the biggest experience bugbears across the customer lifecycle. And it’s gaining a way to finally track and measure service capacity across dealerships in the process.
In December, Suzuki Australia took the wrappers of its ‘Hey Suzi’, an AI-powered voice and web assistant. It’s based on homegrown tech from Contact Harald, an Australian scale-up specialising in contact tracking and AI-to-voice technology that gained a big lift after participating in the AWS Generative AI Accelerator. Hey Suzi is also a first in terms of AI deployments for the automotive company globally.
Using Hey Suzi, customers now have the ability to book their scheduled servicing for their Suzuki vehicles 24/7 over the phone or via the Suzuki Australia website. They gain transparency to their Suzuki dealer’s availability and the ability to choose an appropriate day and time to book their vehicle in thanks to a unified data view across the dealer network of both customer data and service capacity.
“I could see the benefits of having an AI system – it’s always available, which means it’s absolute service,” says Pachota. “You don’t have to wait on the phone to be transferred to the appropriate department or for someone to be available to speak. You can call Suzi at 1 am on 25 December and still be able to book your service. That’s absolute convenience.”
The AI tech is also being used as a conduit for Suzuki Roadside Assistance calls plus other requests for Suzuki Customer Care and Suzuki Dealerships.
Reading the customer feedback
Pachota says the use case for AI through a customer service and retention lens was a no-brainer when he actually paid attention to customer feedback.
“We looked at something that was pretty common in the feedback, which was the availability of people to get on the phone. At dealerships, you have limited windows and business hours through the day; on weekends it’s even more limited and the back-end and operations for autos for service departments,” he says.
“It’s very difficult to find the time during the working day to call the service centre, hoping someone answers, that you don’t get transferred around, and finally getting through to someone to talk to about servicing your vehicle. Then you have to go through the whole identification process to talk about your car.”
However, there was hefty complexity around building such a system for Suzuki’s dealer network. Most significantly, there are up to nine different dealer management systems in place nationally. There was even an accounting book-based service management process still in place in one dealership. All this had to be overcome to implement a fully connected system enabling Suzi to identify the caller as a Suzuki customer and know the history of their vehicle. It took over 18 months, in fact.
“The AI system needs to have information about the fact you’re the owner of the vehicle, your contact number, that the registration coincides with ownership of the vehicle, and to know for example what service is due next,” says Pachota.
“If you call Suzi now, it’ll say Hi, I’m the AI assistant, I understand you want to book in the Suzuki Swift you purchased last year on such and such a day, and it’s due for its 10,000km service. Would you like to do that at the dealership you bought it at? You say yes, I’d like to, and the next question Suzi will ask is if you’d like to book it in on the next available date, which is Tuesday at 7 am, 8 am, 9 am, or Wednesday at 10 am. You can say none work, what about Friday, Suzi will look and say ok Friday I have three times available. It’ll pop the booking in and you get confirmation via text and email to say your vehicle has been booked in for a service.”
In the first month, Pachota’s key measure of success was making sure Hey Suzi was working across new vehicle deliveries. He cites a strike rate of over 90 per cent thus far, along with average booking times of under 2 minutes. From February, Suzuki is kicking off an advertising and communications campaign to build awareness of Hey Suzi Phase 1 across new and existing customers. The second Phase of Hey Suzi capability will coincide with the introduction of a new service campaign in March.
Even without the campaigning, Pachota notes customers have already begun to use Hey Suzi just by locating links on its website. Suzuki sells an average of 20,000 cars per year and Pachota estimates between 300,000 – 500,000 cars are out there on the roads. The company currently has a database covering about 25 years’ worth of data.
We said we have someone – we did a bit of a Trojan horse approach – in Australia that could be an industry first, and we’d like to be the first to do it. And it’s something based on convenience, which is probably what is hurting our retention the most… We obviously had to make a bit of investment to do so, but the dealers got behind us as well saying they’d like to do this. Japan turned around and said go for it. — Michael Pachota, GM, Suzuki Australia
Training, learning, and clean data
With Hey Suzi reducing expensive call centre overheads and ensuring 100 per cent service inquiry coverage, Pachota says dealers were quickly behind the tech rollout. Selling in an AI project to a global team not yet ready to take on the AI-to-voice piece globally also proved relatively straightforward.
“We said we have someone – we did a bit of a Trojan horse approach – in Australia could be an industry first, and we’d like to be the first to do it,” says Pachota. “And it’s something based on convenience, which is probably what is hurting our retention the most… We obviously had to make a bit of investment to do so, but the dealers got behind us as well saying they’d like to do this. Japan turned around and said go for it.”
Critical to any AI rollout is ensuring the data it’s being fed is accurate. In advance of Hey Suzi’s rollout, Suzuki reviewed all data and found common areas where data could be dirty, such as dealers not entering customer data correctly. More broadly, the business has been reviewing data and information collection for the last decade.
“Our data and data retention in terms of customer information went from 55-60 per cent of customer information that wasn’t bogus, to upwards of 95 per cent in the last 5-6 years. Our information collection is healthy. But we did a review to make sure all touchpoints were covered,” Pachota says.
“And we could identify particular dealers not collecting as per the other dealers. Some were new dealers and didn’t understand the process, and there were a few other bits and pieces. But the first thing we did was get field teams to educate those people on the input of data. We created an onboarding document together with Contact Harald. Then we had individual training through DMS providers as well to make sure certain fields were made available.”
A division has been set up to review additional features or connections with customer data to ensure security and privacy. Pachota also points to measures put in place from a cyber security perspective. Suzuki gains access to customer data via the purchase contract signed by customers.
Another critical step was training the AI around available hours for booking hours and service bookings. “There were fields that needed to be in there to tip off for example that there is a 22,000km service, and what does that service entail, how much time do you need to complete that service. That would then trigger a booking space available for 2 hours over here, which is just enough time to do that 22,000km service,” Pachota explains.
“There are all these triggers in the background we had to make sure were available for every customer and the journey of their car, within a five-year span of ownership at first. Then we started going back to customers in the 10-year span.”
Vigorous testing saw the team “trying to break it every single day – that was our mission”. “It was very hard to do it but we found ways and I’m grateful we did. Because those areas can’t be broken now,” adds Pachota.
Measurement and transparency
Capturing 100 per cent of service inquiries was the first goal for Suzuki Australia. “Before having Suzi, we didn’t have a way to understand how many phone calls or inquiries on services were made to dealers. We only had a booking form on the site, which was an email sent to the dealer, but no direct dealer calls,” Pachota says.
“This way we have 100 per cent tracking – it’s the only way to book your service in with us. So the transparency of understanding what’s going on from a service perspective is huge.
“That in itself is gold – it’s already paid for itself. We pay so much money and so many brands pay so much money to get better, cleaner, more accurate data around inquiries, where they go and how it all leads to a sale.”
With such metrics to hand, Suzuki can then measure how many people called versus how many cars were sold that same time 12 months ago and work out the retention rate of people coming back to service, and what years they’re calling in and coming back.
“We’ll be able to also understand the capacity of dealerships. Some tell us they just don’t have the resources in terms of technicians and we rock up and we see 14 Nissan Navara vehicles in there. Now we’ll be able to measure properly,” Pachota says. “It’s not to say to keep the dealers honest, that’s not the case, it’s about a business and partnership to make sure we are doing the right thing for each other and giving each other the best opportunities.”
The huge sense of relief Pachota feels upon finally launching Hey Suzi is palpable. “I was pulling my hair out just wanting to launch it. There were big hurdles but we got over it, and hairy moments. For something that’s first and new to the market, it was always going to be difficult to set up and get everyone onboard,” he says.
Take the small problem of one dealer booking in services in a physical accounting book. “We literally built them a DMS that works specifically for Suzi to plug into the network 100 per cent,” says Pachota.
As to the next steps, the feature is already there for the network to use Suzi to call customers for use cases such as a sales promotion, customer vehicle recalls or service campaigns.
“If there are 500 customers to call, which can take them a week on average to call, Suzi can do that in one day,” Pachota adds. “In the future, it could be from a test drive booking perspective, which we’re working on in the background.”
3 Feb 2025 - Article by Mi3 News Plus