# Patient context, ready for safer handoff. Pokee AI helps care teams turn long clinical histories, nursing notes, vitals, labs, medications, and recent events into a concise, reviewable handoff summary. This demo focuses on healthcare workflow: clearer context, safer transitions, and secure handling of sensitive patient data. **Healthcare focus:** The demo uses synthetic patient data to show how Pokee AI can support nurse handoff, patient-context review, and clinician-supervised summarization in privacy-sensitive care settings. ## Long context Reads many notes, vitals, labs, meds, and events without losing the patient story. ## Secure Designed for protected health data, access controls, and auditable workflows. ## Deployable Supports private cloud, on-prem, and on-device options. ### Pokee Patient Context Demo ### Ready #### MR ### Maria R. · Room 418B Admitted for pneumonia · 68F · Full code · English/Spanish #### High watch **Overnight change:** oxygen increased from 2L to 4L nasal cannula after desaturation to 89%. Fever improved after antibiotics. Potassium is low and replacement is ordered. **Incoming nurse should watch respiratory status, fall risk, and pending chest X-ray.** ### Respiratory risk O2 requirement increased overnight. ### Fall risk Needs assist to bathroom. ### Family context Daughter asks for update after rounds. ### Medication Antibiotic due at 10:00 AM. ## Capabilities ### Designed around healthcare context. A useful nurse handoff assistant needs more than a short summary. It needs to understand long patient context, protect sensitive health data, show where information came from, and fit into hospital deployment requirements. ### Long patient context Review a large window of patient history, clinician notes, medication changes, lab trends, vitals, care plans, and shift events. ### Health-data security Support protected care workflows with access control, minimum-necessary data use, encryption, audit logs, and source-backed outputs. ### Hospital-controlled deployment Deploy inside hospital infrastructure, an enterprise VPC, private cloud, or other controlled environments where patient data must stay governed. ### On-device and edge options Support low-latency clinical workflows closer to the care team, with reduced external data movement where device deployment is appropriate. ### Traceable handoff output Show key claims with source references so nurses can verify the underlying note, vital sign, lab result, medication, or order. ### Clinician review built in Keep AI-generated summaries as drafts until a nurse or clinician reviews, edits, and approves the final handoff context. ## Workflow ### From messy chart data to trusted handoff context. Pokee AI organizes approved clinical data into a structured, reviewable handoff so nurses can quickly see what changed, what is risky, and what needs attention next. 1. ### Ingest approved health data Uses authorized clinical sources such as EHR fields, nursing notes, vitals, labs, medications, orders, and care plans. 2. ### Reason across the chart Finds what changed, what matters, what is missing, and what needs nurse attention across long patient records. 3. ### Draft nurse-ready handoff Produces SBAR summaries, checklists, alerts, and pending-task reminders tailored to shift-change workflow. 4. ### Clinician verifies The nurse or clinician reviews, edits, approves, or rejects the draft before it becomes official. ## Interactive demo ### Shift handoff copilot, on synthetic patients. Select a synthetic patient and switch between SBAR, tasks, and timeline views. ### Patient queue | Patient | Context | |---------|---------| | MR Maria R. | Room 418B · Pneumonia · 68F · Full code | | DK David K. | Room 312A · Post-op hip replacement · 74M · Fall precautions | | AS Amina S. | Room 205C · Heart failure exacerbation · 59F · Fluid restriction | #### Pokee AI handoff summary Oxygen requirement increased overnight from 2L to 4L nasal cannula after desaturation to 89%. Fever improved after antibiotics. Potassium is low and replacement is ordered. Incoming nurse should monitor respiratory status, complete antibiotic administration, maintain fall precautions, and follow up on pending chest X-ray. ### Situation 68-year-old female admitted for pneumonia. Currently on 4L nasal cannula after overnight desaturation. ### Background History of COPD and hypertension. Daughter reports increased confusion compared with baseline. ### Assessment Respiratory status is the main watch item. Potassium is low. Fall risk is elevated due to weakness and bathroom assistance needs. ### Recommendation Reassess oxygen needs, administer antibiotic at 10:00 AM, verify potassium replacement, maintain fall precautions, and follow chest X-ray result. ## Built for the stack you defend ### Privacy-first data handling Use minimum-necessary data, encryption, role-based access, audit logs, and secure integration patterns. ### Clinician verification Generated handoff summaries remain drafts until reviewed by the appropriate nurse or clinician. ### Source-backed handoffs Each key claim can link back to the underlying chart record, note, lab, medication, vital sign, or event. ### Flexible healthcare deployment Support cloud, private cloud, on-prem, and on-device options depending on hospital security requirements. ## Deployment & Governance ### Private deployment A dedicated managed tenant, your own VPC on AWS, Azure or GCP, on premises, or air-gapped. Your data never leaves your boundary. ### Approvals and audit Approval gates wherever you set them, role-based access, and a full audit log. Every agent action is attributable and reviewable. ### Measurable pilots Pilots run on two or three real workflows, typically live inside two weeks. Measure time saved, accuracy and cycle time before rollout.