How Sai works
We’re building Sai to give patients a personal care assistant they can speak to 24/7. It answers patients’ questions about their care plan, guides them through prescribed routines and assists when difficulties arise.
“Hi Sai” is the wake word that starts a voice interaction. The patient then explains what they need in their own words, and Sai replies through the wearable. Its own mobile connection lets patients use it at home and while out, without a smartphone or home broadband.
Sai starts conversations to carry out scheduled check-ins, remind patients about prescribed routines and ask about changes it notices. We’re also developing ways for patients to call family members and request transport or meal delivery. These services require mobile coverage and depend on which services their care provider makes available.
Understanding daily life
The wearable records movement and activity throughout the day and night. Daily check-ins and other conversations capture what the patient reports, alongside speech and sentiment analysis. Sai analyses this information over time to give care teams detailed insights into the patient’s recovery, ongoing health and daily routines.
Sai analyses movement and activity throughout the day and night, tracking the patient’s usual routines and identifying changes in mobility, activity levels and night-time movement. This gives care teams detailed insights into recovery and changes in daily activity between visits, so they can assess progress and decide when a patient needs follow-up.
During daily check-ins and other conversations, patients report what is going well, whether they have completed prescribed routines, any difficulties and what assistance they need. This gives care teams the patient’s own account alongside movement and activity data, so they can assess how the care plan is working and decide what needs attention.
Sai analyses changes in the patient’s speech alongside what they say, comparing each conversation with their usual speech patterns. This gives care teams another way to identify changes in the patient’s condition, even when the patient says nothing is wrong. Combined with daily check-ins and movement data, speech and sentiment analysis provides a more detailed picture of recovery and ongoing health, helping clinicians decide what needs further assessment.
Assessing the need for assistance
Sai analyses the incoming data alongside the patient’s history and care plan. It compares activity with the progress expected in their plan and looks for changes in their usual movement and speech patterns. When that assessment identifies a need for assistance, the platform triggers the appropriate care protocol to manage the response.
Sai asks follow-up questions when a patient’s answer is unclear. If readings from the wearable are missing, Sai makes that clear to the care team. Speech and sentiment analysis adds evidence to the assessment; diagnosis remains the responsibility of a qualified clinician.
The protocol service
A protocol sets out the steps for handling a particular care need, from the event that starts it to the assistance and follow-up it provides. The Protocol Service runs these steps within the platform, selecting the next question or action from the patient’s answers, history and care instructions.
Configuration
A check-in template can be used for medication, hydration or recovery exercises. For a medication check-in, its settings specify when to ask, which questions to use and how to respond to answers such as a reported missed dose or a request for help.
Patient assignment
The assignment supplies the patient’s care instructions, personal settings and responsible contacts. These determine the guidance Sai can give and who should receive a request for additional help.
A medication check-in set for 9 am begins at that time and asks the patient about their medication routine.
A patient is falling behind the activity planned for that part of the day, so the platform starts a recovery check-in to understand the difficulty and offer assistance.
The patient asks for help understanding their exercise instructions, prompting Sai to begin the relevant guidance.
A possible fall can initiate a fall-management response to seek a reply and arrange emergency help, including a 911 call when required. Emergency calling is part of the intended service and requires configured, tested response arrangements.
An illustrative recovery example
The patient has returned home with prescribed exercises. Their assigned recovery protocol enables Sai to explain the routine, encourage them to begin, help with difficulties and check their progress.
The patient’s situation
Sai’s assistance
Sai explains the purpose of the exercises using the patient’s approved recovery information and asks which parts they would like explained further.
“Would it help if I explained what your recovery plan says these exercises are for?”
The protocol in operation
The recovery protocol starts this scheduled conversation and supplies the exercise instructions assigned to the patient. Their answer determines which explanation Sai gives next.
Illustrative intended experience, not evidence of a demonstrated clinical outcome. Physical assistance and clinical assessment remain with qualified people.
Working with your team
When a protocol calls for your team’s help, the request appears in the care team’s dashboard, alongside the patient’s explanation, relevant observations and a record of Sai’s assistance. Staff can review the history, contact the patient and record their response against the request.
The platform tracks who is responsible, whether they have responded and which requests still need attention. If a response is overdue, it follows the escalation steps in the assigned protocol. Sai follows up with the patient to ask whether they received the requested assistance, whether it addressed the difficulty and what further help they need. Their answers update the care team’s record, so staff can review the outcome and decide what to do next.
Sai keeps family members informed about the patient’s daily care, progress and difficulties, even when they live far away. It helps them stay involved in the patient’s care plan, understand when their support is needed and stay connected with the patient.
Learning across patients and providers
Sai brings together patient observations, conversations, care provided and subsequent progress. Across your organisation, this wealth of data can reveal emerging risks, recurring patient needs and opportunities to improve care quality and outcomes. Care teams can identify patients who need attention and escalate concerns, while leaders can use patterns across teams and locations to improve services and plan resources.
Anonymised data across the Sai platform extends that learning beyond your own patients. Comparing care approaches and outcomes across a wider population can reveal patterns that are difficult to see within one organisation, helping you identify opportunities to improve your service and make better-informed decisions about patient care.
Explore what Sai can bring to your customers and your organisation.
Talk to the team →