How David Jones is using digital assistants to improve customer service
- 16 July, 2020 12:34
Managing stakeholder expectations around the artificial intelligence (AI) journey, embracing an iterative approach and building internal capability are key lessons David Jones has learnt after transforming its customer service approach.
Speaking during this week’s Salesforce Live A/NZ event, the department store giant’s general manager of customer service, Daniele Iezzi, detailed how it’s deployed AI and virtual assistants as part of a digitisation of service capabilities internally in the past year. He also highlighted how these capabilities were employed to meet unprecedented customer demand during the peak of the COVID-19 crisis.
The 182-year old company became part of the Country Road group of brands after being acquired by Woolworth’s Holding Group in 2014, and regionalised its customer service teams 12 months ago.
Struggling to keep up with omnichannel customers as a result of outdated systems and processes, the business worked to rejuvenate its in-house customer service capabilities over the past year.
“It’s safe to say the challenges faced by company and retailers in general today are unprecedented as the industry undergoes a structural change,” Iezzi said. “This is more pronounced in the department store sector, as we’re seeing increased competition from international retailers, specialty stores and a shift to online shopping.
“As we pursued a digital-first strategy, online sales as a percentage of company sales grew significantly. This placed pressure on customer service team, driving 50 per cent growth in call volumes with the main inquiry being related to order status. Coupled with outdated systems, the experience wasn’t living up to customers’ expectations.”
Having opted to keep capabilities in-house, David Jones partnered with Deloitte to implement a solution that could future-proof its customer service offering. This saw it deploy Salesforce Service Cloud as well as AWS Connect for telephony. Iezzi said the combination is about empowering agents to resolve customer inquiries by providing the right info at the right time.
Through an AWS Connect AI-based virtual assistant, customers are automatically identified by their mobile phone number upon calling the contact centre, and matched with any existing profile in Salesforce CRM. The assistant then can greet a customer by name. The system also uses natural language to determine the reason for their call.
“If it’s checking order status, the assistant will confirm that in your Salesforce profile, along with your details, then use an API call to our system and delivery partner to relay order status without need for human intervention,” Iezzi explained. “Alternatively, it can transfer you to an agent to speak to a human. This is really powerful, as order status inquiries are our main source of calls.”
The system also automates case creation and pre-populates screen pop-ups with contact details and query types so agents know who is calling and why. Iezzi said the tailored capability has helped improve efficiency, allowing service teams to focus on more complex tasks.
In addition, the system uses identifiers such as customer loyalty status to route a call to a VIP queue. If the 24/7 assistant can’t resolve an issue, it automatically creates case in Service Cloud with contact details plus the reason for the call so a team can action the case in the morning.
Iezzi said initial results are encouraging, with a 35 per cent reduction in average phone call handling. Simple customer intents, such as trading hours and returns policy, are also completely automated.
The biggest challenge for the assistant is when a customer wants to speak to a department in-store. “To direct a call to appropriate extension, the assistant has to understand the name of store, department and the brand customer is looking for,” he noted.
“We have 43,000 extensions, 50 store names, and thousands of brand names, many of which are Italian and French. There’s a way to go but it continues to learn.”
COVID-19 crisis management
The biggest test to date for the AI-powered system has been transitioning the customer service team to work remotely due to COVID-19. Iezzi said online store sales doubled during lockdown, with the team experiencing 500 per cent growth in customer contacts by channel.
“This growth coupled with delays in dispatch and delivery, led to unprecedented demands on our customer service teams,” he said. “At one stage, customers were telephoning every 2 seconds. Wait times ballooned to levels we weren’t comfortable with.”
In response, the David Jones team identified three ways it could improve ability to answer these inquiries. The first was human intervention.
“Despite have the AI assistant, we needed to increase team headcount to deal with inquiries,” Iezzi said. Due to the drop in-store traffic, David Jones redeployed sales assistants into the service team.
These staff members had the benefit of knowledge of customers to bring to this team, he said. What’s more, the AWS Connect system could upscale quickly and work on existing team devices across the country.
The second step was to leverage live chat. Iezzi said David Jones had already increased headcount in this channel to provide another contact option and deflect telephone calls. While results were encouraging, again the team was soon overwhelmed.
To combat this, David Jones implemented a chatbot using dropdown menus to understand customer intent and obtain details like name, case number or product info. This reduced handling time and improved agent capability.
The third step was implementing a callback function to try and cope with wait times of over one hour. This saw AWS Connect automatically making calls when a customer reached the top of the queue, and showed their details to an agent to conduct the call.
“While it wasn’t smooth sailing, we did see some encouraging results, including 58 per cent reduction in time taken to answer calls, and a 20 improvement to abandonment rates,” Iezzi said. “We also had 97 per cent response rates in live chat.”
In addition, up to 50 per cent of calls were handled completely by the virtual assistant. “The combination of Amazon Connect virtual assistant + Salesforce Service Cloud allowed us to keep every channel open as we transitioned to working from home model. This wouldn’t have been possible with previous systems,” Iezzi said.
Eye on the new normal
Having progressed from a period of disruption and consolidation, Iezzi said David Jones now has three priorities. The first is to identify further opportunities to automate against customer intent, deflecting more calls from the service team.
“We’ll continue to gather information and insights with customers, as each contact adds to profile in Salesforce,” Iezzi said. “We’ll continue to monitor the AWS Connect virtual assistant to deliver more seamless experience.”
Secondly, David Jones wants to build on its expertise using the skills and expertise of in-store assistants. “Earlier this year, we commenced proactively selling on live chat, yet out product knowledge sits with in-store teams,” Iezzi said.
“We’re looking to implement chat and messaging in-store, giving customers more options for engagement, and allowing sales assistants to view customer shopping history. This will enable them to sell the right product more easily and efficiently.”
Thirdly, it’s about refining hybrid working models. “Up to this point, our approach to service was quite siloed. By flexible working model, scale up and down more easily, and work to a single view of customer across multiple areas of our business,” Iezzi said.
Through the process, Iezzi said he’s learnt the importance of taking stakeholders on the AI journey.
“It’s important to manage their expectations about what AI can deliver and how quickly it can do so,” he said. This includes customers, too, as “not everyone likes to speak to a robot”, Iezzi said.
Another lesson is the importance of taking incremental steps, especially when using AI. This makes the process less overwhelming for stakeholders, and allows teams to refine issues in a more measured way, he said.
As a final piece of advice, Iezzi said it’s important to build AI capability across your in-house teams. “Don’t forget IT colleagues may be new to this area too,” he added. “This capability will ensure you have support you need so you can continue to focus on what matters most – serving customers.”
Follow CMO on Twitter: @CMOAustralia, take part in the CMO conversation on LinkedIn: CMO ANZ, follow our regular updates via CMO Australia's Linkedin company page, or join us on Facebook: https://www.facebook.com/CMOAustralia.