Liquor retailer achieves $2.9M revenue uplift with AI Shopping Assistant

Leading liquor retailer sets sights on innovation

As one of Australia’s leading liquor retailers, the company provides customers with an accessible range of alcoholic beverages across a vast national network of stores and its ecommerce channels. The retailer, who has been in partnership with Probe CX for 9 years, continues to invest in and innovate its online shopping experience. 

Capitalising on changing buyer behaviour

Cost of living pressures have changed the way alcohol is enjoyed in Australia. A reduced propensity to drink and dine out has seen consumers turning to discount liquor retailers to help recreate the experience at home. And as shoppers increasingly purchase online, an opportunity presented itself for the company to capitalise on changing consumer behaviour. 

The challenge? Maximise website conversion rates and revenue per online session, while emulating the personalised experience of shopping with an in-store assistant. This sought to enhance the customer experience and increase basket value, without increasing live chat support costs. 

‘It was a combination of seeing how we could help customers, but also the client being very open to collaboration, experimentation and innovation,’ recounts Chaz Domingo, Senior Management Consultant at Probe Group, who worked with the retailer’s customer success team on the project. 

They wanted to test whether enhancing the existing chatbot, primarily used for FAQs, would aid product discovery and conversion, and subsequently influence revenue generation. 

The goal

Improve ecommerce website conversion and increase average basket size to create a meaningful uplift in revenue. 

A better experience from discovery to checkout 

To proactively engage visitors, an AI-powered Shopping Assistant was designed in collaboration with stakeholders and deployed on the brand’s ecommerce site. The Assistant uses a proprietary Virtual Agent powered by a complex Natural Language Understanding (NLU) model (the AI component) that understands context and responds to user queries — even complex requests. 

Probe CX first mapped the customer journey and defined buyers’ considerations when shopping online for alcoholic beverages, ensuring the Assistant would effectively manage a range of queries. 

‘There was a lot of discussion around the criteria that would be important to customers to make it a good user experience,’ says Chaz, with category and type of drink, occasion and price point deemed the salient criteria. This helped shape early prototypes. 

‘It was an iterative process of testing to see which version would best fit the client’s needs,’ she adds. 

The live Shopping Assistant enables intent-driven personalisation of relevant products, whether for a gift, occasion, food pairing or something else. For example: a user seeking wine pairing recommendations for seafood answers a series of questions and receives tailored recommendations to shop. This makes high-value SKUs more discoverable, streamlines the path from discovery to checkout, and increases order value. 

Additionally, the Assistant is significantly more cost effective than assigning human agents to the task (though agent handoff is enabled as a fallback option), without compromising the experience. 

The customer success team gained full visibility into success via a Business Intelligence (BI) dashboard integrated with Google Analytics, building confidence in the solution. It tracks user interactions with the Shopping Assistant and attributes revenue when users convert thereafter. The dashboard insights also offer valuable opportunities for future decision-making and optimisation.

Significant revenue uplift achieved 

The Shopping Assistant delivered massive, quantifiable conversion and revenue success. 

‘Currently the conversion rate of the Shopping Assistant is consistently two to three times higher than users who don’t engage,’ reports Chaz, noting that seemingly small gains translate into significant revenue growth for a retailer of this size.

Case in point: over $357,000 in measurable sales revenue was generated within a defined period by the Shopping Assistant, alongside resolving more than 3,700 queries without human intervention and delivering running cost savings of over $3,600.

To date, the Shopping Assistant has: 

This liquor retailer continues to optimise the Shopping Assistant with Probe CX and explore the ways in which it can support revenue generation into the future.