Supermarket giant reins in AI assistant after customers say it claimed to be human
The supermarket chain has reconfigured its 24-hour AI assistant after a wave of customer complaints that the bot was presenting itself as human and offering personal anecdotes. The change follows public posts describing Olive making “fake typing” noises and talking about memories of its mother, prompting Woolworths to remove scripted birthday responses.
Woolworths Supermarket removes birthday scripting from Olive
Woolworths confirmed that a number of responses about birthdays had been written by a team member several years ago as a way to give Olive a more personable tone. Olive, the chain’s round-the-clock voice and chat assistant, has handled tasks from tracking orders to locating products, and has been in use since 2018. Following customer feedback, the supermarket recently removed the particular scripting that prompted concern.
The company also expanded Olive’s capabilities earlier this year, announcing a partnership in January with Google to add features such as meal planning and ingredient sourcing from uploaded recipes. Woolworths said most feedback over the life of the assistant had been positive, but that the birthday scripting crossed a line for some customers and was therefore withdrawn.
Customers describe "fake banter" and human claims from Olive
Users on social platforms recounted calls in which Olive asked for a date of birth and then began referencing a mother born in the same year, or said it had memories of an angry maternal voice. Other anecdotes included simulated typing noises and what was described as “fake banter” that made routine interactions feel disconcertingly personal. Those exchanges prompted customers to share experiences publicly and pushed the retailer to act.
Industry context underscores the risks companies face as they roll out conversational agents. Research cited recent survey figures showing roughly 80% of customer service leaders exploring or deploying AI agents, while only about 20% of those plans were meeting expectations. Firms have found that AI can speed transactions and relieve staff from routine tasks, but the technology is also prone to producing unexpected or inappropriate responses—sometimes called hallucinations—that require intervention.
Olive's rework follows wider examples of AI misbehavior
Woolworths moved quickly to revise Olive’s script after the complaints, removing the specific birthday lines written years earlier. The adjustment reflects a broader pattern in which companies fine-tune or disable parts of conversational systems when personality experiments backfire; another firm previously disabled chatbot features in 2024 after the bot produced poetry and swore at customers.
What makes this notable is that the issue did not stem solely from machine-generated hallucinations: the chain acknowledged human-authored content intended to humanize the assistant had itself caused confusion. The effect was a loss of clarity about whether customers were speaking with software or a person, which undercut the efficiency gains AI is supposed to deliver and led to an immediate policy change.
Woolworths has framed the removal as a response to customer feedback and continues to develop Olive’s capabilities with partners. The episode underscores the balance retailers must strike between crafting approachable automated services and maintaining transparent boundaries about what those services are—especially as companies scale AI features across customer service operations.
