Automation

Stroke survivor forced to switch GP after AI receptionist fails

A Doncaster woman's experience highlights accessibility gaps in healthcare automation as practices adopt AI phone systems.

Omega Editorial· August 26, 2026· 3 min read

Accessibility concerns surface as GP practices deploy AI phone systems

A 71-year-old stroke survivor was forced to register with a new GP practice after an AI receptionist system repeatedly failed to understand her speech, according to the BBC.

Judith Butterfield attempted to book appointments at the ASA surgery in Armthorpe, near Doncaster, five times following her stroke in January. The facility had deployed EMMA, a nationwide AI reception system designed to reduce phone wait times. Each attempt ended in failure when the system couldn't process her speech, which had been affected by the stroke.

"There was a point in the conversation where I had to repeat my name and she just couldn't understand me," Butterfield told the BBC. "If it's a person they can say, 'could you spell that?' But the call just dropped off. It was very difficult and upsetting."

The system's requirement that patients not use speakerphone created an additional barrier. Butterfield's stroke had primarily affected the right side of her body, making it impossible to hold the phone in one hand while operating the keypad with the other.

Why it matters

As healthcare providers rush to adopt AI automation to manage staffing shortages and patient volume, this case exposes a critical gap: systems optimized for typical speech patterns may systematically exclude patients with disabilities. The incident raises questions about whether AI deployment in healthcare settings adequately considers accessibility requirements or undergoes testing with diverse patient populations before rollout.

Practice reverses course after patient complaint

The ASA surgery has since decommissioned the EMMA service. Practice manager Claire Oxley confirmed the system is no longer in use and said the facility has recruited additional reception staff to improve service.

Butterfield has registered with the Burns Practice in Doncaster, where human receptionists handle calls. "I was amazed when I spoke to a person," she said.

According to Healthwatch, the healthcare watchdog, Butterfield's experience is not isolated. Other patients in Yorkshire have reported similar difficulties with AI reception services.

Company offers solution after the fact

QuantumLoopAI, the company behind EMMA, said every practice using the system can configure "priority routing" to allow certain patients to reach human staff on every call. However, this feature apparently was not activated for Butterfield or communicated to her during her attempts to book appointments.

A company spokesperson said they would welcome the opportunity to investigate what happened with Butterfield's consent and "help put it right."

The case was first reported by the BBC.

#healthcare ai#accessibility#speech recognition#patient experience#uk healthcare#disability rights

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Automation

Automation· 3 min read

WhatsApp Commerce Drives $530M in Revenue Across 317 Brands

Meta partner Zefir releases benchmark data showing 88% average open rates and double-digit ROI across customer lifecycle automation.

Via Automation Watch · Sep 24, 2026
Automation· 3 min read

Five Million Industrial Robots Now Operating Worldwide

China accounts for 59% of global deployments as Asia drives adoption, while Europe lags and the U.S. overtakes Japan in market size.

Via Automation Watch · Sep 24, 2026
Automation· 4 min read

Simbe's 3,000 Tally Robots Mark Shift to Public-Space Automation

The retail inventory robot's commercial milestone signals robotics moving beyond controlled factory floors into unpredictable human environments.

Via Automation Watch · Sep 24, 2026