The problem
Are you a registered manager, operations director or owner of a care provider?
Beecared4, a Salford home care provider supporting around 250 clients with 92 care professionals, faced the pressures every provider knows: tight margins and rising wages.
A one-hour assessment meeting could be followed by up to six hours of typing to produce the care plan.
What they did
The consultation is securely recorded and transcribed. The system suggests answers for the routine parts of the assessment.
High-risk clinical questions get no AI suggestion at all. The care manager reviews and confirms each answer before anything is saved.
What was reported
The registered manager describes being highly hesitant at first, and changing her mind once she understood how the system handled ethics and personalisation.
Where people stay in control
No suggestions on high-risk questions
Clinical risk questions are answered by the care manager, with no AI suggestion offered.
Nothing saved without review
The care manager confirms each answer before it becomes part of the record.
Lessons for care providers
- Decide in writing where AI may never go. Name the questions and decisions that stay entirely with people, before choosing any tool.
- Protect the conversation. Use AI to reduce typing after the visit, not to shorten time with the person.
- Make review the default. Every suggested answer is confirmed by a manager before it is saved.
- Read supplier figures carefully. Ask for the method behind any time saving and measure it in your own service.
Our view
The most important design decision here is a refusal: no AI suggestions on high-risk clinical questions. Start your own project by writing that list.