Woolworths Q&A: Using Conversational AI to Personalize and Enhance Customer Experience


Customer happiness and satisfaction is priority number one at Australian-based Woolworths. “Creating an excellent customer experience is the reason we exist,” noted Alex Colorado, Engineering Manager and Chapter Lead for the Conversational AI squad at WooliesX. “… All of our journeys start by looking at the customer problem we’re trying to solve. Once this is clear, our team of designers, developers and AI trainers work together to create a seamless experience. By working this way, we delight the customer by creating something easy and intuitive; and support our contact center teams by allowing them to focus on more complex customer queries.”

I recently spoke with Alex to learn about this customer-centric approach — and how Woolworths plans to use digital and artificial intelligence (AI) to further improve the customer experience (CX).

What do you do in your current role?

Alex Colorado: I help design and build the backbone of how we enable customers to interact with us conversationally through our Customer Hub contact center.

What were you using before implementing a Genesys solution to manage your CX?

Colorado: We have been a long-term Genesys customer. Our original IVRs used traditional DTMF. However, we have integrated the Genesys platform with Google Dialogflow, giving us conversational capability, which powers our smart IVRs.

Tell me about your current customer experience setup.

Colorado: We have around 800 agents based in Australia and New Zealand. The biggest group of agents support our eCommerce online shopping business, which has grown rapidly over the last few years. The contact center also covers other Woolworths Group businesses, like Insurance, Rewards and Mobile.

Our Genesys platform is mainly an on-premises solution, with specific components running both on-prem and cloud solutions. We are looking to move to Genesys Cloud soon.

How has your approach to CX shifted over the past five years and why?

Colorado: If we go back a few years, there was an industry push to make speech IVRs obsolete and a return to traditional DTMF IVRs. This was probably due to speech recognition technology and its technical limitations at the time, which resulted in poor customer experiences.

Conversational AI technologies and the advent of deep and machine learning are now setting the future for customer experience and contact centers. With these new technologies, we’ve been able to move IVRs from being business-focused — interact with us how we think you should — to being customer-focused — interact with us in whatever way feels natural.

How do you think your CX infrastructure or strategy will evolve over the next three to five years?

Colorado: For the vast majority of people, having a conversation is the most natural form of communication. It’s because of this that I see virtual assistants aiding the customer experience in everything we do.

Over the next three to five years, we will see a significant shift from virtual assistants just answering simple FAQ-styled questions to solving real-world problems whenever the customer needs them.

Think of this, for example: looking in your cupboard and adding grocery items by voice, asking your Woolworths virtual assistant to come up with meals, or getting them to help you navigate the fast way through a shopping center.

Our strategy is to be at the forefront of this technology. This means trying new things early, failing quickly, learning from those failures (and successes), and applying them to everything we do.

What challenges did the COVID-19 pandemic pose to your customer service and contact centers? And how did you solve them?

Colorado: Woolworths is the largest food retailer in Australia. At the beginning of the pandemic, Australia went into lockdown. Because of this, we saw a massive shift to online shopping. This change in customer behavior, in turn, created rapid growth in interactions with our contact centers as many customers engaged with us digitally for the first time.

Our eCommerce virtual assistant, Olive, was vital to help our contact center manage the influx of calls and chats. Olive was able to keep customers up to date on COVID-19-related announcements and provide multiple self-service options.

Olive handled thousands of customer interactions end to end — meaning customers received the information they needed in a timely manner and didn’t need to speak to a customer service representative. This has provided significant relief to our contact centers at a time of exceptionally high demand.

Where do most of your customer interactions happen?

Colorado: There has been significant growth in chat interactions, especially since our virtual assistants have been integrated with more pages on our websites and there’s been a push to get them featured on our Woolworths mobile apps. However, when it comes to speaking to a human and having in-depth interactions with our customers, voice (IVR) is still our number one channel.

What channels do you currently support or plan to support in the future?

Colorado: Thanks to the way our platform was designed, we have been able to integrate seamlessly with multiple channels, like web chat, mobile app, IVR, Google Assistant, Google Business Messages and Facebook.

One channel that I’m pretty excited about is Google Chat (formally Google Hangouts), which we’re using for our internal chatbot. With this, we’ve integrated everything through Genesys, which means customers don’t need to use a chat embedded on a webpage to ask for help. They can use Google Chat from their Gmail inbox.

How are you using conversational AI for your CX goals?

Colorado: The first way we learn to communicate is via speech. For us, it’s about taking the most natural way to interact with one another and use that to help solve customer problems. It’s also worth noting that conversational AI has lots of benefits in terms of accessibility. There is a lot of potential and upside to this technology — and we’ll continue to look for ways to apply it to solve problems for our customers.

If your customers have self-service options, what does that journey look like for them?

Colorado: When we started building bots at Woolies, we started with simple FAQ bots with simple self-service options. Since then, we have been pushing our bots to be more contextualized and personalized.

As an example of a personalized conversation, customers could say something like, “When I received my delivery, the eggs were broken.”

And, based on the user’s phone number, our virtual assistant, Olive, would:

  1. Look up their latest order
  2. Check they have eggs
  3. Note down the reason for the refund (broken item)
  4. Confirm if there was anything else wrong with their order
  5. Process the refund and send a confirmation email to the customer

In the not-too-distant future, Olive could also offer the option to redispatch certain items back to the customer automatically.

Other examples of self-service options include:

  • In-store product finder
  • Track my order
  • Store locator
  • Password reset
  • Real-time store busy-ness
  • ID and verification
  • Provide feedback

What are the biggest opportunities for digital and AI within CX?

Colorado: It’s still early days, and there are multiple areas that present opportunities to improve the customer experience. One thing that we’re exploring is: How do we use data to anticipate what the customer is calling or chatting with us about?

An example of this is, using various data points that can tell us a lot more about the call reason, such as:

  • Does the customer have an order in flight?
  • Has an order just been delivered?
  • What have customers called us about before?

Using this data, we could anticipate what a user is calling us about and guide them down the right path. Similarly, we could work with agents to adjust the conversational tone to be more sympathetic or fun, depending on how they’ve reacted to us in the past.

To learn more, read the Woolworths case study.