AGL taps AI agents in retail transformation

AGL is starting to make use of AI agents as part of its retail transformation, with the initial focus on "employee-based use cases".

Bebin Thomas (left) and Sian Wesley (right) speak on stage.

The retail transformation sees AGL switching its customer relationship management system from SAP to Salesforce, and adopting low-code process automation tooling as well.

At an Agentforce summit in Melbourne late last month, it was revealed that AGL is starting to experiment with Agentforce - an AI agent platform made by Salesforce - in a “testbed” environment.

The early use cases focus on improving the ability of customer service staff to do their jobs, freeing up time to provide advice to customers.

“For us, it was how can we trial [Agentforce] with a small group of users, switch on the out-of-the-box capability and see what benefits we get,” customer operations transformation owner Sian Wesley said.

“For the users we’ve switched it on for, we’ve got automatic voice transcription and then case summarisation [use cases enabled]. 

“We were able to turn that on, get some feedback, [and] make some tweaks.

“For those [customer service] agents using it, they don’t have to write contact notes at the end of the call, so the benefit for them was really instantaneous, and that provides a really good opportunity for us to think about what’s the next evolution for us.”

Wesley said that at a strategic level, the ambition for Agentforce and AI generally is to free up time for customer service advisors to engage customers around emerging renewable energy products and services.

“[We’re] really looking at how we might use Agentforce to get [advisors] to have answers for customers faster,” Wesley said.

“One of the things we’re looking at this side of Christmas is can we plug Agentforce into knowledge management materials, so rather than someone having to go and find the answer to what they need to know, they can just go and query [it] to service that customer better.”

In the longer term, use cases for the technology could be to assist in product servicing and the personalisation of energy offers.

Additionally, AGL may look at some customer-facing use cases for the technology.

“I think in the short term our view for Agentforce is getting some kind of employee-based use cases first before we would go and put that in front of customers, so making sure that we really understand the [Agentforce] technology,” Wesley said.

Head of architecture for customer markets Bebin Thomas added that AGL’s use of Agentforce had been assisted through preparatory “IA” work, which he described as both information and integration architecture.

“We have been intentional with our mission in terms of building a robust information architecture as well as an integration architecture just so we can seamlessly integrate Salesforce with other applications within the business ecosystem, so that we can provide via our customers and the agents with actionable insights,” Thomas added.