Samio Deploys Room-Level Hardware to Automate Clinical Records

Ai4

Software-only AI scribes plateau because they rely on fallible clinician memory and manual app inputs, making passive room-level spatial hardware essential for autonomous documentation and compliance tracking.

AI-assisted summary of Presentation by Samio Systems — original source. Check the original for full context.

Source published: 2026-08-17

Detailed notes

## Background Published on YouTube by Ai4 on August 17, 2026, this presentation features Dr. Mohamad Elzaim, Co-Founder of Samio Systems, introducing a hardware-enabled spatial intelligence platform for healthcare environments. The talk outlines how room-level spatial infrastructure can eliminate provider documentation burdens, enable autonomous clinical workflows, and establish real-time operational awareness across health systems.

## TL;DR Dr. Mohammad Elzaim introduces Samio Systems' spatial clinical intelligence platform, which pairs wall-mounted room anchors, passive staff badges, and LiDAR mapping to replace manual software AI scribes with zero-friction autonomous capture. The system preserves multi-role intake dialogues, delivers bidirectional in-room bedside translation, and maintains an objective spatial audit trail to evaluate emergency responses like Code Blues and verify nursing compliance. Beyond documentation, the platform analyzes care burden and staff path burden to optimize hospital operations, while previewing future features like voice-activated smart call lights and personalized patient discharge apps.

## Atmosphere & style The presentation carries a pragmatic, authoritative, and clinically grounded tone, driven by Dr. Elzaim's perspective as an active primary care physician who intimately understands administrative burnout. It transitions smoothly from critiquing the friction of existing software scribes to presenting concrete, high-stakes clinical scenarios like malpractice defense and resuscitation timelines. The delivery feels crisp, compelling, and operationally rigorous, offering an engaging vision of how physical ambient infrastructure can transform modern healthcare delivery.

## Key takeaways - [00:04](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=4s) Current healthcare AI models rely on EMR data derived from fallible clinician memory after long shifts, while software-only AI scribes struggle with usability friction and low adoption rates. - [02:53](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=173s) Samio introduces spatial clinical intelligence using wall-mounted anchor hubs, passive staff badges, and LiDAR spatial mapping to capture clinical workflows autonomously without manual data entry. - [05:23](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=323s) Presence-based hardware triggers eliminate the cognitive tax of manually operating software scribes, allowing facilities to achieve 100% staff adoption without requiring provider behavior change. - [08:21](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=501s) Ambiently recording medical assistant intake preserves detailed patient history, dynamically guides intake questions, and allows physicians to review summaries and place pre-orders before entering the exam room. - [10:37](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=637s) Embedding bidirectional real-time translation into in-room hardware removes the delays of shared interpreter carts, ending the hazardous rationing of translation services during routine care. - [12:18](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=738s) Objective spatial timelines verify staff bedside presence during emergencies and legal investigations, disproving abandonment claims while highlighting systemic issues such as shift-change coverage gaps. - [15:02](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=902s) Passive room logging quickly resolves regulatory investigations in skilled nursing facilities while simultaneously enforcing accountability to mandatory patient rounding protocols. - [16:19](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=979s) Care Burden analytics track cumulative direct bedside care minutes across clinical roles, facilitating acuity-balanced nurse staffing and providing early warning signals for patient deterioration. - [16:19](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=979s) Path Burden analytics map physical transit patterns and bottlenecks across floor plans, enabling administrators to optimize storage layouts and redirect wasted steps into patient care. - [18:22](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=1102s) Future extensions leverage existing room hardware to offer personalized multilingual discharge summaries and voice-activated smart call lights that classify intent and route requests intelligently.

## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=4s) Introduction and the Foundation of Healthcare AI

Dr. Mohammad Elzaim introduces himself as a primary care physician practicing in Dallas, Texas, and the CEO and co-founder of Samio Systems. He acknowledges his team members in the audience, including co-founder Dr. Ahmed Abousaid and lead engineer Jonathan Grimm. Setting the stage for the talk, Dr. Elzaim presents what he terms "spatial clinical intelligence"—a paradigm that moves beyond standard AI software by embedding intelligent technology directly into the physical clinical environment.

```mermaid flowchart TD accTitle: The Unreliable Memory Pipeline of Current Healthcare AI subgraph Capture["Original Clinical Interaction"] A["Patient Encounter"] --> B["Clinician Memory\n(Post 50-patient day / 12-hr shift)"] end subgraph Storage["Medical Record"] B --> C["EMR Data Entry\n(Prone to omission & error)"] end subgraph Downstream["Healthcare AI"] C --> D["Existing AI Models\n(Summarizers, workflow tools)"] end ```

He prompts the audience to reconsider the entire data foundation of modern healthcare. While the Electronic Medical Record (EMR) sits unquestioned at the center of every system—with current AI tools built entirely around summarizing, reorganizing, and extracting data from it—the raw input originates almost entirely from human memory. Clinicians, nurses, and medical aides routinely write notes after seeing a cluster of patients or late at the end of grueling shifts. When AI models train on records generated from the tired memories of overworked healthcare professionals handling 50-patient days or 12-hour shifts, the underlying data is inherently unreliable. Existing software AI scribes attempt to intervene, but their daily adoption rates plateau between 25% and 40%, burdened by usability friction. Samio’s philosophy flips this paradigm: rather than creating another tool to write the chart, let the clinical environment itself capture and document care in real time.

---

[02:53](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=173s) The Architecture of Spatial Clinical Intelligence

To make clinical spaces aware and responsive, Samio relies on a lightweight two-component hardware setup paired with rapid spatial mapping. The primary component is the Samio anchor device—a compact blue unit mounted to the wall in every exam or patient room, functioning as that room's central intelligence hub. The second piece is a passive clinical badge clipped to existing ID holders worn by physicians, nurses, and medical assistants. Staff do not need to interact with the badge other than recharging it when the battery is low.

```mermaid flowchart LR accTitle: Samio System Setup and Components subgraph Deployment["Room Setup"] A["Anchor Device\n(Wall-mounted hub)"] B["iOS App LiDAR Scan\n(Generates 2D blueprint)"] end subgraph Tracking["Active Space"] A --> C["Intelligent Exam Room"] B --> C D["Passive Staff Badge\n(No button presses)"] --> C end subgraph Output["Output System"] C --> E["Integrated RTLS +\nAmbient Intelligence"] end ```

Dr. Elzaim emphasizes that installation is designed for simplicity and self-deployment. Facilities order devices corresponding to their room count, and any handy person can hang an anchor like a picture frame and plug it into a standard wall outlet. Using an iPhone or iPad equipped with built-in LiDAR sensors, the Samio setup app quickly scans the clinic to generate a precise 2D blueprint. Combining the 2D layout with the anchor hubs and staff badges forms a Real-Time Location System (RTLS). While traditional RTLS hardware exists, legacy systems are expensive, notoriously difficult to install, and fail to integrate with ambient documentation pipelines.

- **Anchor Device**: Wall-mounted hub installed in each exam room, plugging into standard power. - **Passive Staff Badge**: Lightweight wearable requiring zero manual button clicks or interactions during clinical care. - **LiDAR Setup App**: Scans physical clinic layouts using Apple LiDAR to create accurate 2D spatial blueprints.

---

[05:23](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=323s) The Fatal Flaw of Software AI Scribes vs. Autonomous Capture

Software-based AI scribes require a continuous mental tax from clinicians. A provider must carry a smartphone or laptop, select the right patient, input encounter types, obtain transcription consent, position the microphone, and remember to tap start, pause, and stop. In chaotic environments like emergency departments or urgent care clinics—where doctors are routinely interrupted, redirected to urgent cases, or pulled into hallway consults—manual capture inevitably breaks down. Providers forget to unpause, record discussions with colleagues while previous patient sessions remain open, or select the wrong chart, directly corrupting the transcript.

Samio eliminates manual operation by using physical presence as the trigger: walking into the room starts the encounter, and walking out stops it. Notes and role-specific summaries populate without touching a screen or an app. Dr. Elzaim points out the adoption benefits of removing friction, noting that instead of begging a veteran clinician—such as a 78-year-old physician accustomed to paper charts—to navigate a new mobile app, administrators simply clip a badge onto their coat. The notes appear automatically, enabling clinics to achieve systemic, facility-wide efficiency gains rather than relying on piecemeal individual adoption.

---

[08:21](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=501s) Multi-Role Intake and Predictive Clinical History

Patients frequently complain about having to repeat their medical history multiple times during a single doctor's visit. Dr. Elzaim explains that a clinical visit does not begin when the physician enters; it begins during rooming, when a medical assistant (MA) or nurse conducts intake. This initial intake captures the freshest, most detailed context regarding symptom timelines, medication changes, and allergy details. However, because standard intake workflows compress complex stories into just two or three summary words in the EMR, massive data loss occurs, forcing the physician to re-interview the patient from scratch.

```mermaid flowchart TD accTitle: Multi-Role Intake and Physician Preparation Flow subgraph Intake["Rooming Phase"] A["MA Enters Room with Badge"] --> B["Anchor Records Full Intake Context"] B --> C["AI Intake Guidance & Suggested Questions"] end subgraph Preparation["Pre-Visit Review"] C --> D["Semi-Detailed Summary Formed"] D --> E["Physician Reviews Notes & Places Pre-Orders"] end subgraph Exam["Physician Encounter"] E --> F["Doctor Enters to Confirm, Expand, or Correct"] end ```

By equipping MAs with badges, the ambient anchor records the entire intake discussion, preserving nuanced details. As the MA conducts rooming, the system can dynamically guide the dialogue by suggesting appropriate follow-up questions tied to the patient's chief complaint. Before the physician ever steps foot inside the room, they receive a comprehensive, structured intake summary. Clinicians can review the narrative, confirm specifics, make necessary corrections, and even enter preliminary lab or imaging orders in advance—streamlining the encounter so doctors can focus on high-level clinical decision-making.

- Preserves rich initial intake data that is otherwise condensed into minimal EMR fragments. - AI dynamically suggests targeted follow-up questions to clinical support staff during rooming. - Equips physicians with structured background information to place preliminary orders prior to room entry.

[10:37](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=637s) Real-Time Multilingual Translation at the Bedside

By automating documentation, Samio frees clinicians from spending excessive time gathering historical details, allowing them to redirect their energy toward clinical decision-making and patient communication. Rather than being tethered to computer screens and keyboards, providers can give their full attention to the human sitting in front of them. Building on transcription, the platform introduces ambient real-time translation to address a chronic operational friction point in multilingual care settings.

```mermaid flowchart TD accTitle: Legacy vs Samio Translation Workflow subgraph Legacy["Legacy Translation Process"] A1["Locate Shared iPad Cart"] --> A2["Wheel to Room & Connect Service"] A2 --> A3["Interpreter 5-Min Opening Spiel"] A3 --> A4["Rationed Use: Major Encounters Only"] end subgraph Samio["Samio Ambient Translation"] B1["Hardware Device in Room"] --> B2["Provider Speaks Native Language"] B2 --> B3["Speaker Broadcasts Translated Audio"] B3 --> B4["Patient Replies in Native Language"] B4 --> B5["Instant Doctor Translation"] end ```

In typical hospitals and clinics, language interpretation relies on scarce shared equipment—often a single mobile interpreter iPad per floor. Clinicians must track down the cart, wheel it into the room, dial into an interpretation service, and wait for a remote translator to deliver a standard five-minute introductory spiel before the clinical conversation can even begin. Because this setup often takes longer than the actual clinical encounter, staff inevitably begin rationing translation, reserving it only for high-stakes interactions like new diagnoses, surgical consents, or formal physician rounds. This rationing introduces substantial clinical risk.

Samio eliminates this friction by embedding translation directly into the room's permanent hardware. The same wall-mounted device handling transcription acts as an active bidirectional interpreter. When a doctor speaks, the device’s integrated speaker outputs the speech in the patient's native language; when the patient responds, it immediately plays back the translation in the doctor's language, requiring zero physical setup or third-party dialing.

- **Hardware integration**: Translation runs directly through the in-room speaker and microphone array without requiring dedicated translation carts or iPads. - **Workflow impact**: Eliminates translation rationing across routine nursing checks, medication passes, and ad-hoc clinical assessments.

---

[12:18](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=738s) Spatial Intelligence in Action: The Code Blue Scenario and Objective Look-Back

To demonstrate the power of room-level spatial intelligence, a real-world clinical scenario illustrates how the system records and interprets emergencies. In the case of JM, a 67-year-old male admitted to Room 101 with pneumonia, the patient was initially placed on low-flow oxygen. Over four hours, his respiratory status progressively deteriorated into cardiopulmonary arrest, triggering a hospital-wide Code Blue.

```mermaid flowchart LR accTitle: Code Blue Timeline and Look-Back Verification subgraph Event["Room 101 Emergency"] E1["Gradual Respiratory Decline"] --> E2["Cardiopulmonary Arrest"] E2 --> E3["Code Blue Broadcast"] end subgraph SpatialData["Spatial Intelligence Log"] E3 --> S1["Real-Time Team Convergence"] S1 --> S2["Retrospective Entry Log: 9 Badge Check-ins"] S2 --> S3["80 Bedside Minutes Documented"] S2 --> S4["47-Minute Shift Gap Identified"] end ```

When a Code Blue is initiated, the platform’s spatial mapping displays staff converging on the specific room in real time, logging exactly who arrives, their clinical role, and their precise duration at the bedside. Beyond tracking the active resuscitation, the system provides a retrospective look-back over the preceding hours, generating an objective timeline of every room entry and physical evaluation.

This room-level record proves critical when adverse outcomes lead to legal challenges. In JM's case, the family filed a malpractice lawsuit alleging the patient was neglected and left unmonitored for hours prior to the arrest. Because healthcare facilities cannot place video cameras inside patient rooms for privacy reasons, resolving such disputes traditionally relies on piecing together subjective provider notes and electronic medical record (EMR) timestamps. Samio's spatial audit trail objectively refuted the abandonment claim:

- **Documented bedside presence**: 9 badge-verified room entries totaling roughly 80 minutes of direct bedside care across four hours, involving nurses, physicians, and respiratory therapists. - **Operational findings**: The look-back uncovered a 47-minute unmonitored window during nursing shift change. - **Quality improvement**: The platform provides defensible documentation against unfounded liability claims while isolating systemic operational lapses—such as shift-change coverage gaps—that leadership can directly remediate.

---

[15:02](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=902s) Resolving Skilled Nursing Allegations and Enforcing Rounding Compliance

The necessity of an objective spatial audit trail is equally prominent in long-term care and skilled nursing facilities (SNFs). Recurring patient complaints frequently trigger mandatory state regulatory investigations, requiring administrators to spend days pulling physical charts, interviewing staff, and attempting to reconstruct timelines from fallible human memory.

In one highlighted example from a skilled nursing facility, a resident submitted a formal complaint alleging that an individual named Michael entered their room in the middle of the night and physically assaulted them. Utilizing the spatial timeline, administrators reviewed every badge entry into that specific room across the entire night shift. The system verified that no unauthorized individuals entered the room, resolving an investigation that typically consumes multiple days in a matter of minutes.

Crucially, this objective visibility "cuts both ways." While examining the timeline to investigate the assault allegation, the record revealed that the assigned Certified Nursing Assistant (CNA) was only entering the room once every four hours. Because facility policy mandated Q2 rounding (checking on the resident every two hours), the system simultaneously exposed an active failure in care compliance. By providing an unalterable, honest record, the infrastructure protects facilities against exaggerated or unfounded claims while holding internal staff accountable to clinical care standards.

---

[16:19](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=979s) Advanced Workflow Analytics: Care Burden and Path Burden

The third core capability of the platform extends beyond individual patient encounters into enterprise workflow analytics. Historically, hospital operational metrics—such as true patient dwell time, bedside nursing duration, waiting room delays, and room utilization rates—have relied on coarse estimates and guesswork. Spatial intelligence transforms these into precise, continuous metrics categorized into clinical care burden and physical path burden.

```mermaid flowchart TD accTitle: Care Burden and Path Burden Applications subgraph Analytics["Spatial Intelligence Analytics"] CB["Care Burden Analytics"] PB["Path Burden Analytics"] end subgraph CBOutcomes["Care Burden Applications"] CB --> CB1["Measure Direct Bedside Minutes per Role"] CB --> CB2["Acuity-Based Nurse Staffing Balancing"] CB --> CB3["Early Deterioration Signal Tracking"] end subgraph PBOutcomes["Path Burden Applications"] PB --> PB1["Map Staff Transit Heatmaps"] PB --> PB2["Identify Repetitive Supply Room Trips"] PB --> PB3["Optimize Facility Layouts & Storage"] end ```

**Care Burden** measures the exact direct care time a patient receives across an entire shift, broken down by clinical role (physician, registered nurse, technician, respiratory therapist). - **Workload rebalancing**: Rather than assigning nurses purely by patient headcounts, charge nurses can balance assignments based on real-time accumulated care burden, ensuring no single nurse is overwhelmed by disproportionately demanding patients. - **Early clinical warning**: Tracking a patient's care burden curve over time provides a predictive clinical signal. When direct care requirements climb steadily above baseline, it frequently serves as the earliest indicator of physiological deterioration.

**Path Burden** visualizes the physical movement patterns of clinical staff across a floor plan. Movement tracking reveals operational bottlenecks, such as a nurse making repetitive, cross-unit trips back and forth to a distant supply room throughout a shift. By highlighting transit time that detracts from direct patient care, hospital management can execute targeted quality improvement initiatives—such as relocating high-frequency supply storage closer to active patient clusters—to convert wasted transit steps back into bedside care minutes.

---

[18:22](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=1102s) Future Horizons: Patient-Facing App and Contextual Smart Call Light

The presentation concludes with an overview of two forward-looking products in active development, both utilizing the pre-installed in-room hardware infrastructure without requiring additional devices.

```mermaid flowchart LR accTitle: Voice-Activated Smart Call Light Routing A["Patient Speaks: 'Hey Samio...'"] --> B["In-Room Ambient Microphone"] B --> C["Natural Language Classification"] C --> D1["Mobility Assistance -> CNA"] C --> D2["Pain Medication -> Assigned RN"] C --> D3["Dizziness Warning -> Urgent Care Team"] ```

The first initiative is a dedicated **Patient-Facing Application** centered on personalized discharge education and transparent operations: - **Tailored after-visit summaries**: Automatically generates post-encounter summaries derived directly from the doctor-patient dialogue rather than dispensing generic handouts. Delivered in the patient’s primary language, summaries include specific medication guidance, return precautions, and self-care protocols for home reference. - **Real-time spatial wait times**: Integrates live spatial room tracking to provide patients in waiting areas with accurate, dynamic wait-time estimates, reducing front-desk inquiries and waiting-room frustration.

The second initiative is a voice-activated **Smart Call Light** that replaces conventional uninformative push-buttons: - **Natural language intent classification**: Patients activate the system verbally by stating specific needs (e.g., *"Hey Samio, I need help getting up"*, *"Hey Samio, I need my pain medications"*, or *"Hey Samio, I feel dizzy"*). - **Intelligent routing with context**: The system parses the clinical urgency and intent, immediately dispatching the request to the appropriate team member with detailed situational context. - **Accessibility**: Provides immediate, hands-free call access for motor-impaired, post-surgical, or immobilized patients unable to reach or physically depress a legacy call pendant.

## Notable quotes > What if clinical documentation and workflow intelligence happened automatically, without any reliance on provider behavior? — Dr. Mohammad Elzaim

> [00:04](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=4s) spatial clinical intelligence — Dr. Mohammad Elzaim

> [00:04](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=4s) rather than creating another tool to write the chart, let the clinical environment itself capture and document care in real time. — Dr. Mohammad Elzaim

> [05:23](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=323s) babysitting technology — Dr. Mohammad Elzaim

> [05:23](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=323s) walking into the room starts the encounter, and walking out stops it. — Dr. Mohammad Elzaim

> [08:21](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=501s) a clinical visit does not begin when the physician enters; it begins during rooming, when a medical assistant (MA) or nurse conducts intake. — Dr. Mohammad Elzaim

> [15:02](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=902s) cuts both ways — Dr. Mohammad Elzaim

> [18:22](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=1102s) Hey Samio, I need help getting up

> [18:22](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=1102s) Hey Samio, I need my pain medications

> [18:22](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=1102s) Hey Samio, I feel dizzy

## Mentioned Dr. Mohammad Elzaim, Dr. Ahmed Abousaid, Jonathan Grimm, Samio Systems, Ai4, Apple LiDAR, iPad, iPhone, Electronic Medical Record (EMR), Real-Time Location System (RTLS), JM, Michael

## Video description Mohamad Elzaim - Co-Founder, Samio Systems

What if clinical documentation and workflow intelligence happened automatically, without any reliance on provider behavior? In today’s healthcare environment, healthcare personnel are overwhelmed by administrative burden, with nearly half of their time consumed by documentation, creating inefficiencies and contributing to burnout. In this talk, I will introduce SAMIO Systems and explain how embedding spatial intelligence directly into the clinical environment—through a hardware-enabled, room-level system—fundamentally shifts how data is captured and workflows are executed. I will illustrate how this approach enables autonomous documentation, real-time operational awareness, and system-wide standardization that software-only solutions cannot achieve. Attendees will gain insight into how this infrastructure creates scalable “forced efficiency” and unlocks a new layer of clinical and operational intelligence. If you’re looking to understand where healthcare is heading beyond traditional AI tools, this session will offer a compelling look at what comes next.

[Source](https://www.youtube.com/watch?v=rVeXKuK2IHQ)

中文简报

Samio部署病房级硬件实现临床记录自动化纯软件AI记录助手陷入发展瓶颈，因其依赖容易出错的人工记忆和应用端手动输入，这使得被动式房间级空间硬件成为实现自动化文档记录与合规追踪的关键。 Mohammad Elzaim博士介绍了Samio Systems的空间临床智能平台。该平台结合壁挂式房间定位锚点、被动式员工胸牌与LiDAR空间建图技术，以零摩擦的自主采集取代需手动操作的软件AI记录工具。系统不仅能完整记录多角色接诊对话、提供病房床旁双向翻译，还能保留客观的空间审计线索，用于评估Code Blue等紧急抢救响应并核验护理合规性。除文档记录外，该平台还通过分析护理负担与医务人员动线负荷来优化医院运营，同时展望了语音激活智能呼叫灯及个性化患者出院App等未来功能。 Samio Systems 推出空间临床智能平台，通过壁挂式锚点硬件、无源员工工牌与 LiDAR 空间建模实现基于物理在场触发的自主病历记录，打破传统纯软件 AI 书记员 25% 至 40% 的采纳率瓶颈，达成全院 100% 的无感应用。 (5:23) 系统在医护助手接诊环节即可完整记录详尽病史并动态引导问询，同时在室内硬件中直接嵌入双向实时翻译功能，终结了传统共用翻译手推车带来的延误与日常护理中的翻译配给难题。 (10:37) 客观空间时间线能精确记录床旁人员进出与抢救过程，在法律调查中提供客观证据以驳回疏忽指控并暴露交接班盲区，同时帮助专业护理机构迅速应对监管指控并核查定时巡房合规性。 (12:18) 照护负担（Care Burden）分析功能可按临床角色精确统计床旁直护时长，既能辅助排班主管实现病情严重度平衡的护士调配，也能通过照护时间激增作为患者病情恶化的早期预警信号。 (16:19)

Reading summary

Dr. Mohammad Elzaim, CEO and co-founder of Samio Systems, argues that software-only healthcare AI fails because it relies on fallible clinician memory and introduces usability friction. He proposes 'spatial clinical intelligence'—embedding passive ambient hardware directly into clinical rooms to automate documentation, translation, and operational tracking. Flaws in Current Healthcare AI and Memory-Based Records (0:04) Dr. Mohammad Elzaim contends that modern healthcare AI relies on a fundamentally compromised foundation: the Electronic Medical Record (EMR). Because clinicians, nurses, and medical aides routinely enter notes late after 12-hour shifts or 50-patient days, downstream AI models are trained on records distorted by fatigue and omitted details. Software AI scribes attempt to capture encounters in real time, but Elzaim notes that daily adoption typically plateaus between 25% and 40% due to operational friction. Software tools require clinicians to juggle devices, confirm consent, switch encounter types, and manually manage start, pause, and stop functions. In high-interruption environments like urgent care clinics and emergency departments, clinicians frequently forget to unpause recordings, capture hallway consults inappropriately, or select incorrect charts, corrupting transcript accuracy. Passive Spatial Hardware and Multi-Role Intake (2:53) To eliminate manual interaction, Samio pairs a wall-mounted anchor unit in each room with passive clinical badges worn on existing staff ID holders. The physical layout is mapped using Apple LiDAR sensors on an iOS device to generate a 2D clinic blueprint, establishing an integrated Real-Time Location System (RTLS). Walking into a room automatically initiates the session, and exiting stops it, requiring zero button presses from clinicians and enabling clinic-wide adoption even among technology-averse staff. This setup also captures the multi-role intake process often handled by medical assistants (MAs) during rooming. Because standard practice compresses detailed initial histories into brief EMR notes, physicians normally must re-interview patients. By recording the MA intake, Samio suggests dynamic follow-up questions during rooming and provides physicians with structured summaries and narrative context so they can confirm details and enter preliminary orders prior to stepping into the room. Ambient Bedside Translation and Objective Spatial Audits (10:37) Samio embeds bidirectional real-time translation directly into the room's anchor unit, bypassing the logistical friction of locating shared translation iPads and waiting through introductory protocols. By conducting translation immediately through the in-room speaker and microphone array, clinics can prevent staff from rationing translation services strictly to major clinical events, thereby protecting routine check-ins and medication passes from miscommunication. Furthermore, passive room tracking produces an objective physical timeline of bedside care. In a featured malpractice case involving a fatal cardiopulmonary arrest (Code Blue) in a 67-year-old pneumonia patient, spatial logs disproved family allegations of total abandonment by confirming nine badge-verified check-ins totaling roughly 80 direct bedside minutes, while simultaneously identifying a 47-minute unmonitored window during shift change. In a skilled nursing facility, an identical audit trail quickly cleared a staff member of an alleged middle-of-the-night assault while revealing that the assigned nursing assistant failed to meet the required two-hour rounding mandate. Workflow Analytics and Forthcoming Patient Interfaces (16:19) Beyond documentation, spatial data converts enterprise hospital metrics from coarse estimates into two granular analytical measures: 'care burden' and 'path burden.' Care burden tallies direct bedside care time categorized by clinical role, serving as both an acuity-based tool for balancing nurse workloads and a predictive signal where sudden increases in care time reflect patient physiological decline. Path burden maps physical staff transit across facility floor plans, revealing operational bottlenecks such as frequent trips to distant supply closets so hospitals can adjust layouts and restore transit time to bedside care. Samio is also developing two patient-facing applications that utilize the existing room hardware without additional devices. A patient application generates discharge summaries in the patient's native language based on the actual doctor-patient dialogue and provides dynamic wait times. Additionally, a voice-activated smart call light system uses natural language processing to categorize requests—such as mobility help, pain medication, or acute dizziness—and route them with clinical context directly to the appropriate staff member. Samio positions ambient spatial intelligence as an infrastructure shift from software documentation tools to environmental clinical awareness. However, facility deployment still depends on physical hardware installation in every room and continuous staff badge compliance across shifts.

精炼阅读摘要

Samio Systems 联合创始人 Mohammad Elzaim 博士指出，当前医疗人工智能过度依赖基于医护疲劳记忆录入的电子病历，导致数据基础存在先天缺陷。他提出“空间临床智能”架构，通过墙面锚点、被动工牌与空间映射，将临床环境转变为实时、无感的全自动数据捕获与运营分析系统。硬件赋能空间智能：消除软件转录阻力 (2:53) Elzaim 认为，传统基于软件的 AI 医疗记录工具要求医护人员频繁手动选择病历、开启暂停录音及调整麦克风，在急诊室或快节奏诊所等易被打断的高压场景中极易失效，导致全院日常采纳率通常停滞在 25% 至 40% 的区间。针对这一痛点，Samio 采用了轻量化硬件架构：在每个诊室安装插入标准电源插座的墙面锚点设备，医护人员随身佩戴免操作的被动感应工牌，并利用搭载 LiDAR 的 iOS 设备快速扫描生成诊所 2D 平面图，从而低成本构建起集成实时定位系统（RTLS）的智能诊室。该系统以医护人员的物理进出作为唯一的数据采集触发机制，跨入房间即可自动激活记录，离开时即刻中止，完全免除了屏幕交互与应用操作负担。Elzaim 表示，这种零心智负担的设计使习惯纸质病历的高龄资深临床医生也能无缝接入，从而推动医疗机构实现 100% 的全院级普及率，无需依赖碎片化的个人使用意愿。多角色分诊协同与床旁即时双向翻译 (8:21) 在传统门诊中，患者常抱怨需向不同医护重复陈述病史。Elzaim 解释称，临床接诊始于医疗助理（MA）的分诊室测量，但这一阶段复杂的病情叙述往往被压缩为 EMR 中的寥寥数语，造成严重信息损耗，迫使医生进房后从头问起。Samio 系统通过让分诊助理佩戴工牌，实现对初始沟通细节的完整留存，并由 AI 针对主诉动态推荐后续提问。在医生步入诊室之前，系统已生成结构化分诊摘要，供医生提前审核并预开检查或检验单，提升接诊效能。此外，针对跨语言诊疗的阻力，该平台利用固定在墙面的锚点硬件提供床旁即时双向翻译。传统模式下各楼层共享翻译 iPad 推车，不仅需要费时寻找与连接，还迫使医护对日常查房或给药进行“翻译配给”。Samio 的固定式麦克风和扬声器阵列能够自动在医生和患者各自的母语间进行实时双向语音播报，无需额外移动设备或第三方呼叫等待，消除了常规护理中的沟通风险。客观空间轨迹化解医疗纠纷与合规监管 (12:18) 由于患者隐私限制，病房内无法安装视频监控，医疗争议往往依赖于主观陈述和病历时间戳。Elzaim 通过一名因肺炎入院并转为心肺骤停（Code Blue）的 67 岁患者案例说明，家属指控院方存在长达数小时的照护疏忽，而系统的空间审计日志显示，在病情恶化前的 4 小时内共有 9 次工牌验证的进房记录，累计直接床旁照护时长约为 80 分钟，客观反驳了家属的弃管指控；同时，该数据亦精准定位出护士交接班期间出现的 47 分钟未监测窗口，为院方提供了针对性改进依据。类似的空间可追溯性在长期护理与专业护理机构（SNF）中同样发挥了“双刃剑”效应。在应对一起住户指控夜间遭外人闯入袭击的调查中，系统在几分钟内通过门禁轨迹排除了外部人员进入的可能性，终结了耗时数天的常规调查；然而，回溯记录同时发现负责的认证护理助理（CNA）每 4 小时才进入房间一次，违背了机构每 2 小时巡视一次的强制规定。客观数据在保护机构免受不实指控的同时，也严格追究了内部的违规行为。照护与动线分析：重塑机构运营指标 (16:19) 医院运营长期依赖对患者停留时间、床旁照护耗时和房间周转率的粗略估计。Samio 将空间智能拓展至两项核心运营分析指标：“照护负荷”（Care Burden）与“动线负荷”（Path Burden）。其中，照护负荷能够细致量化每名患者在整个班次内接受医生、护士、呼吸治疗师等各角色直接床旁照护的精确分钟数。护士长可依据动态累积的照护负荷科学均衡分配人力，避免单一护士承担过载负荷；此外，若患者的照护负荷曲线持续反常上升，常被视为生理指标恶化的早期临床预警信号。动线负荷则通过对医护人员在物理楼层中的移动热力图进行追踪，量化往返穿行消耗的时间。例如，系统能识别出护士在班次中频繁跨区往返偏远库房取用耗材的瓶颈，促使管理层将高频物资重新部署至患者集中区域，切实将无效步行时间转化为直接床旁照护时间。延伸规划：动态患者端应用与智能呼叫铃 (18:22) 基于既有的室内硬件底座，Samio 正在研发两款无需额外采购设备的延伸产品。首先是患者端移动应用，该应用直接基于实际医患对话自动生成多语言的定制化出院后随访摘要，提供精确的用药指引与复诊注意事项，替代传统的通用纸质材料；同时，它整合了实时的空间排队追踪，向候诊室患者推送动态等待时间，以缓解焦虑情绪并减少前台咨询。另一项规划为语音激活的智能呼叫铃系统，旨在取代传统功能单一的物理按压按钮。患者通过自然语言（例如“Samio，我需要翻身协助”、“我需要止痛药”或“我感到头晕”）发出请求，系统自动识别其意图与临床紧急程度，精准派发任务至对应的 CNA 或主管护士，并附带现场情境信息。此举不仅提升了调配响应效率，也为行动不便或术后无法触及传统呼叫按钮的患者提供了无障碍求助渠道。 Dr. Elzaim 总结认为，空间临床智能通过将物理环境升级为客观感知平台，摆脱了传统医疗对医护记忆和手动软件录入的依赖。然而，系统在多人员重叠环境下的语音分离准确度，以及大规模跨院区部署时的硬件维护成本与数据融合深度，仍需更广泛的实践检验。

Detailed notes — zh

## Background 本内容由 YouTube 频道 Ai4 于 2026 年 8 月 17 日发布，主讲人为 Samio Systems 联合创始人兼 CEO Mohamad Elzaim 博士。视频主要介绍了 Samio Systems 如何通过在诊疗环境中部署结合硬件的“空间临床智能”（Spatial Clinical Intelligence），摆脱对医护人员手动操作的依赖，实现全流程自动病历记录、实时运营感知与系统级标准化。 ## TL;DR Elzaim 博士指出当前医疗 AI 严重依赖医护人员在疲惫状态下凭记忆录入的失真 EMR 数据，且传统软件类 AI 书记员存在操作繁琐、采纳率低等痛点。为此，Samio Systems 推出了由“房间锚点设备”和“被动式临床工牌”组成的空间智能体系，实现医护进屋即触发的零接触自动化记录、接诊信息无缝流转及全天候房间级实时翻译。此外，该系统还能在 Code Blue 急救和护理机构纠纷中提供精准的时间线回溯以规避法律风险，并通过量化“照护负荷”与“动线负荷”深度优化排班和医院物理布局，未来还将拓展至患者端 App 与语音智能呼叫铃。 ## Atmosphere & style 整个演讲兼具一线临床医生的实战敏锐度与科技创业者的严谨理性。主讲人以极具共情力的方式直击医护人员在繁琐文档与多语言沟通中的疲惫痛点，随后通过真实的 Code Blue 诉讼回溯和护理院调查案例，展现出空间数据对于医疗抗辩与质控改进的硬核威力。现场节奏明快、条理清晰，既有对现有技术局限的尖锐洞察，又洋溢着以人为本、将技术“隐形化”以彻底解放临床生产力的笃定与前瞻性。 ## Key takeaways - [00:04](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=4s) 当今医疗 AI 工具高度依赖 EMR 数据，但 EMR 数据本质上源于高压、过度劳累的医护人员凭记忆补写录入，存在天然的失真与不可靠性。 - [00:04](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=4s) 传统软件类 AI 书记员工具依赖医护主动操作，机构采纳率仅在 25% 至 40% 之间，难以推动全院级别的标准化效率提升。 - [00:04](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=4s) Samio 提出的“空间临床智能”（Spatial Clinical Intelligence）理念旨在将记录功能下沉至物理空间，由诊疗环境在信息产生的源头自动感知与捕获。 - [02:53](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=173s) 系统通过墙面锚点设备、被动式临床工牌以及 iPad LiDAR 快速建模构建，提供极简自主部署的低成本高精度实时定位与环境感知能力。 - [05:23](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=323s) “物理存在即触发”彻底消除了传统 AI 软件对医护人员注意力份额的持续占用，避免了因人为忘记启停导致的转录遗漏与错误。 - [08:21](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=501s) 捕获涵盖分诊与医疗助理（MA）接诊在内的全流程信息，不仅能通过 AI 实时引导问诊，还能提前生成病史摘要，消除患者重复陈述病情的痛点并前置临床决策。 - [10:37](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=637s) 房间硬件内置麦克风与扬声器，提供全天候全病房的双向实时语音翻译，打破了传统共享 iPad 翻译模式带来的资源配给与沟通壁垒。 - [12:18](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=738s) 空间智能生成的时间线能客观记录 Code Blue 等急救事件的到达与停留时长，既能作为关键法律证据有力抗辩不实指控，也能揭示交接班管理漏洞。 - [15:01](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=901s) 在护理机构中，客观出入记录兼具为医护人员洗清虐待虚假指控与监督落实规范查房频次（如 Q2 查房）的双向管理效能。 - [16:24](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=984s) 通过统计床旁照护时长建立“照护负荷”（Care Burden）指标，既可支持更公平的人力排班，又能将负荷持续攀升作为患者病情恶化的早期预警信号。 - [16:24](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=984s) 基于医护人员物理移动轨迹分析“动线负荷”（Path Burden），可定位往返耗材室等低效动线，指导医院空间重构以减少无效奔波。 - [18:23](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=1103s) 系统正进一步拓展患者端个性化就诊摘要（AVS）、高精度排队等候预估，以及复用现有硬件实现语义精准分流的智能语音呼叫铃。 ## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=4s) 医疗 AI 的核心困境：建立在“人类记忆”之上的 EMR 数据作为来自德克萨斯州达拉斯的初级保健医生、Samio Systems 联合创始人兼 CEO，Mohamad Elzaim 博士开场介绍了坐在后排的联合创始人兼 CEO Ahmed Abousaid 博士与首席工程师 Jonathan Grimm。Elzaim 指出，当前的医疗环境过度依赖电子病历系统（EMR），现有的医疗 AI 工具无一不是构建在 EMR 之上，负责总结病历、提取与重组临床数据、自动化行政工作。然而，当追溯这些 EMR 数据的源头时，会发现它们全部来自于医生、护士或医疗助理等医护人员的“记忆”。医护人员通常在接诊多位患者后才抽空写病历，或者索性拖到一天工作结束时统一记录。Elzaim 抛出一个尖锐的问题：在一位医生看完了 50 位患者的疲惫状态下，或是一名护士经历了长达 12 小时的筋疲力竭轮班后，凭记忆手写录入的数据到底有多可靠？答案是“非常不可靠”。当今最先进的医疗 AI 模型，本质上是在使用由高压、忙碌且过度劳累的医护人员凭记忆手工录入的数据进行训练。现有的 AI 书记员（AI Scribe）虽然有所帮助，但其日常采用率仅在 25% 至 40% 之间，且伴随着自身的缺陷。为此，Samio Systems 提出了“空间临床智能”（Spatial Clinical Intelligence）这一全新理念：不再开发又一个仅仅帮忙写病历的工具，而是让整个“诊疗环境”本身来完成记录，在临床信息产生的第一时间于物理空间中直接捕捉。 ```mermaid flowchart TD accTitle: Traditional EMR AI versus Spatial Clinical Intelligence subgraph Traditional["传统医疗 AI 模式：数据失真链路"] A["医护人员接诊患者"] --> B["繁重轮班后凭记忆补写"] B --> C["录入 EMR 系统"] C --> D["传统医疗 AI 工具提取与总结"] end subgraph Spatial["Samio 空间临床智能模式"] E["医护人员进入物理诊室"] --> F["环境自动感知与实时多角色拾音"] F --> G["零操作生成高保真结构化记录与流转"] end ``` ### [02:53](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=173s) 硬件与环境感知：Samio 锚点设备、被动式工牌与极简部署空间临床智能的核心在于让诊疗空间理解其物理环境、正在发生的诊疗行为、参与人员以及停留时长。Samio 系统的落地依赖于两大核心硬件组件： 1. **Samio 锚点设备（Anchor Device）**：安装在每个检查室或病房墙壁上的蓝色方块设备，充当该房间的中央智能枢纽。 2. **被动式临床工牌（Passive Clinical Badge）**：由医生、护士、医疗助理（MA）等医护人员佩戴，只需夹在现有的工牌夹上即可。除了在电量不足时充电外，日常使用中处于完全被动状态，不需要任何人工交互。系统的部署流程被设计为完全自主安装（Self-installable）。诊疗机构只需根据房间数量接收设备，由后勤人员像挂相框一样将锚点设备固定在任何墙面的固定位置并插上电源。随后，操作人员使用专用的 Samio Setup App，通过 iPhone 或 iPad 内置的 LiDAR 激光雷达扫描诊室，即可自动生成诊疗空间的二维平面图蓝图。结合工牌与锚点设备，系统能够构建起高精度的实时定位系统（RTLS）。尽管市场上已存在传统的 RTLS 方案，但普遍面临安装复杂、组件繁多、成本高昂且缺乏深度集成的问题，而 Samio 将空间定位与多模态交互直接无缝融合。 ### [05:23](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=323s) 传统 AI 书记员的致命缺陷与“零接触”自动化听写当前主流的软件类 AI 书记员产品要求医护人员执行繁琐的手动操作流程：携带笔记本或手机进屋、手动输入患者姓名与就诊类型、获取录音知情同意、调整设备摆放角度、手动点击开始录音、就诊结束后手动点击停止。Elzaim 指出，这些操作单独看似简单，但其“致命缺陷”（Fatal Flaw）在于它们需要持续占据医护人员的注意力份额（Attention Slice）。在繁忙急躁的急诊室、诊所或急救中心，医生在不同房间与走廊之间频繁穿梭，极易出现人为疏漏：忘记点击开始或结束录音、选错患者姓名、离开房间折返时忘记暂停、或在讨论其他患者病情时仍未关闭麦克风。每一次微小的人类疏忽都会直接破坏转录质量。医护人员的精力应当 100% 聚焦在面前的患者身上，而不是去“伺候技术”（Babysitting the technology）。 Samio 彻底实现了该工作流的自动化： - **物理存在即触发（Physical Presence Trigger）**：只要医护人员走进房间，系统依靠空间智能感知自动启动录音，离开即结束，无须掏出任何设备或点击屏幕。 - **消除技术采纳阻力**：对于在医院工作数十年、习惯纸质病历的 78 岁老医生 Dr. Jones，不再需要强迫他学习复杂的手机软件，只需将 Samio 工牌夹给他，系统便会自动在他的病历系统中生成高质量笔记。 - **全院级效率提升**：传统的软件工具依赖医生个体自愿使用，效益上限受限于 25%~40% 的自愿采纳率；而 Samio 通过为全体员工配发工牌，直接实现全院范围（Facility-wide）100% 的效率普及。 | 评估维度 | 传统软件类 AI 书记员 (Software-based AI Scribe) | Samio 空间智能系统 (Samio Spatial Intelligence) | |:--- |:--- |:--- | | **设备交互** | 需携带手机/电脑，手动启停、选患者、调角度 | 完全被动式工牌，进出房间物理感知自动触发 | | **注意力占用** | 分散医生对患者的注意力，极易因遗忘导致转录错误 | 零认知负荷，医护人员全神贯注于患者本身 | | **机构采纳率** | 依赖个体意愿，平均仅 25% - 40% | 配发工牌即可覆盖，实现 100% 机构级全员落地 | | **数据采集范围** | 仅限于医生与患者交谈的单一片段 | 覆盖分诊、MA 问诊、医生检查的全流程多角色环节 | ### [08:21](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=501s) 破解“患者重复陈述”痛点与 AI 实时引导问诊在就医过程中，患者最常见的抱怨之一就是“在不同医护面前重复讲自己的病情无数次”。Elzaim 解释，诊疗并非始于医生进门的那一刻，而是始于医疗助理（MA）或护士接诊（Rooming / Intake）的环节。患者在此阶段提供的信息往往最详尽、最新鲜，包括症状严重度、用药变化时间线、过敏筛查等细微线索。然而在现有流程中，患者向 MA 倾诉的长篇详情，最终录入 EMR 的往往只有简陋的“两三个单词”，造成严重的数据流失（Data Loss）。医生进入房间后由于缺乏细节，不得不要求患者将整个病史从头再讲一遍。通过 Samio 系统，接诊的完整过程均被环境智能捕获： - **无缝信息流转**：当医疗助理在房间内完成问诊后，在医生开门进入之前，系统就已经自动提炼并生成了半详细的病史记录。医生进门前便能快速审阅、确认，进门后只需针对性地展开深入评估或修正，彻底消除冗余问诊。 - **AI 实时问诊引导**：系统不仅仅是被动捕获历史，还能在医疗助理接诊的同时，根据患者的主诉（Chief Complaint）实时推荐跟进追问的问题，使基层人员采集到的病史比常规问诊更加完整。 - **决策前置**：医生在进入诊室之前，就已经掌握了足够完备的信息，甚至能够提前开具初步的检查或医嘱，真正践行“医生（MD）的核心价值在于做出临床决策（Making Decisions）”。 ### [10:37](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=637s) 消除语言壁垒：全天候房间级实时翻译在临床工作中，医生往往将大量精力耗费在病史采集与病历录入上。Samio 系统的核心价值在于解放医生的生产力，让他们“减少收集病史的时间，将更多时间用于做出临床决策，并向患者解释沟通这些决定”，真正将目光从冰冷的屏幕和键盘移开，重新聚焦在患者身上。演讲者重点介绍了自己最喜爱的核心功能——**实时多语言翻译**。当前绝大多数医院和诊所都面临多语言患者群体的沟通挑战，但现行的翻译流程极其低效：整层病区通常只配备一台安装了翻译软件的专用 iPad，医护人员必须满楼层寻找设备、推入病房、拨通远程翻译热线，再等待翻译员上线并听完冗长的前置说明。一套流程下来，准备时间往往比实际问诊时间还要长。这导致医护人员被迫“配给使用”翻译资源，仅在告知新诊断、签署手术知情同意书等关键环节才去借用设备，从而埋下了严重的临床沟通风险。 Samio 彻底消除了这一痛点。用于转录问诊的房间硬件设备本身就内置了扬声器与麦克风，支持全天候实时的双向互译。医生用母语与患者交流，设备即时转换为患者的母语播报；患者以母语回应，设备再即时翻译给医生听，全程无需推车、无需拨号，随时可用。 - **传统翻译模式痛点**：单层楼仅一台共享 iPad、设备寻找与接入耗时长、翻译资源被动配给（仅限关键环节）、潜藏医疗隐患。 - **Samio 解决方案**：房间级硬件全时在线、零等待即时双向语音翻译、覆盖每一次日常医患互动。 --- ### [12:18](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=738s) 空间智能实战：Code Blue 急救回溯与医疗纠纷抗辩演讲者通过一个真实的临床案例展示了 Samio 在突发状况下的表现。患者 JM 是一位 67 岁的男性，因肺炎入住 101 病房。入院初期仅接受低流量吸氧，但在接下来的 4 小时内其肺功能急剧恶化，最终发生心肺骤停，医院紧急拉响了“Code Blue”（代码蓝，指患者呼吸心跳停止的最高级别医疗紧急状态）。 ```mermaid flowchart TD accTitle: Samio 空间智能事件回溯与纠纷抗辩流程 subgraph S1["事件发生与数据采集"] A["101 病房突发 Code Blue"] --> B["工牌感应自动记录医护人员进出"] B --> C["沉淀空间轨迹与床旁停留时长"] end subgraph S2["事后调查与抗辩验证"] C --> D{"患者家属起诉\n指控数小时无人看护"} D --> E["客观数据抗辩：\n4 小时内 9 次进出 / 80 分钟床旁照护"] D --> F["运营质控发现：\n交接班期间存在 47 分钟照护空白"] end E --> G["依法驳回不实指控\n保护医护合法权益"] F --> H["定位流程漏洞\n推动医院质控改进"] ``` 在医院模拟演练中，一旦呼叫代码蓝，多位医护人员瞬间涌入 101 病房。依托空间智能技术，Samio 能实时记录每一位进入病房的医护人员身份、到达时间以及停留时长。更关键的是它的“事后回溯能力”：医疗场景出于隐私保护通常无法安装摄像头，当不良预后发生时，“当时到底发生了什么”往往成为家属、医院与医护人员争论的焦点。在该案例中，家属提起诉讼，指控医院在患者骤停前的数小时内将其弃之不顾。Samio 调取的时间线提供了不容置辩的客观房间级数据： 1. **有力抗辩**：在事发前的 4 小时内，共有 9 次经工牌验证的医护进房记录，涵盖护士、医生与呼吸治疗师，累计床旁照护时间接近 80 分钟，彻底推翻了“无人看护”的指控。 2. **暴露隐患**：时间线同时揭示在医护交接班期间，存在长达 47 分钟的查房空白期。正如演讲者所强调的：“Samio 在为既有医疗行为提供有力抗辩的同时，也揭示了真实的运营管理漏洞。”这不仅保护了医院免受不实指控，也为质控部门提供了可落地的改进方向。 --- ### [15:01](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=901s) 护理机构核查：双向还原客观真实的照护记录演讲者分享了自己此前担任两家专业护理机构（Skilled Nursing Facility, SNF）医疗总监时的亲身经历。在其中一家机构，曾有一名常住患者隔月就会提出一次虐待指控。按照规定，州监管部门每次都必须派员调查，而管理层则需要耗费数天时间调取病历、挨个问话，整个调查完全依赖医护人员模糊的记忆和纸质记录。在一次指控中，该患者声称“一名叫 Michael 的男子在半夜闯入房间殴打了他”。通过 Samio 的历史回溯系统，调出房间级时间线后，便能精确还原当晚进出该房间的所有人员名单、进入时间、频次和停留时长。以往耗时数天的调查在几分钟内即可通过客观数据得出结论。这种客观记录的威力在于“双刃剑”效应： - **保护合规操作**：在面对夸大或虚假指控时，以客观数据为医护人员洗清嫌疑。 - **落实流程问责**：在同一份回溯记录中，系统发现护士助理（CNA）并未按照机构规定的“Q2 查房制”（每 2 小时查房一次）执行，实际间隔拉长到了每 4 小时一次。这种客观真实的记录既能抵御外界虚假指责，又能倒逼机构内部规范照护标准。 --- ### [16:24](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=984s) 深度工作流分析：照护负荷与动线优化除了服务于单次诊疗与事件回溯，Samio 体系的第三大核心板块是**工作流深度分析（Analytics）**。系统能够量化以往无法测量的基础指标：患者与医生的真实沟通时长、护士床旁护理时间、患者在诊室的等待时间以及病房的实际利用率。在此基础上，系统通过数据挖掘解决临床运营中的瓶颈问题： ```mermaid flowchart LR accTitle: 空间分析在人员与动线优化中的应用 subgraph Burden["照护负荷分析 (Care Burden)"] A["量化护士每班床旁分钟数"] --> B["按真实负荷均衡排班\n而非仅看患者人头数"] C["追踪单人照护负荷走势"] --> D["负荷持续攀升 ➔ 临床恶化早期预警"] end subgraph Path["动线负荷分析 (Path Burden)"] E["追踪医护物理移动轨迹"] --> F["发现护士频繁往返耗材室"] F --> G["优化耗材室物理布局 ➔ 减少无效走动"] end ``` - **照护负荷（Care Burden）**：有史以来第一次，管理者能够按岗位角色精确统计患者在一个轮班周期内获得的实际床旁护理时长。 - **排班均衡**：当系统识别出某位护士承担了不成比例的超额工作量时，可依据“实际照护负荷”而非粗放的“患者床位数”来合理分配人力。 - **临床早期预警**：连续追踪单个患者的照护负荷曲线，若其所需的护理时间呈持续上升趋势，这往往是患者病情发生潜在恶化最早期、最敏锐的信号。 - **动线负荷（Path Burden）**：通过追踪医护人员在院内的物理移动轨迹（热力线图），管理者发现护士在整个轮班期间频繁往返于病房与位于角落的耗材室。通过分析这些无效走动消耗的时间，医院可以对耗材室进行就近搬迁或重新布局，将护士从奔波赶路中解脱出来，把时间还给床旁患者。 --- ### [18:23](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=1103s) 面向未来的拓展：患者端应用与语音智能呼叫铃在演讲尾声，团队透露了正在研发的两款延伸产品： ```mermaid flowchart TD accTitle: Samio 拓展产品架构 subgraph App["患者端 App"] A1["个性化就诊后摘要\n基于实际问诊生成/母语交付"] A2["高精度实时排队等候预估\n基于空间临床智能数据"] end subgraph Voice["智能语音呼叫铃"] B1["基于现有硬件语音唤醒\n'Hey Samio...'"] B2["意图理解与智能分流\n下床/止痛药/头晕 -> 精准呼叫对应人员"] end ``` 1. **患者端移动应用（Patient-Facing App）**： - **个性化健康宣教**：彻底摒弃通用的模板传单，直接基于该次就诊的实际问诊对话，以患者首选语言自动生成个性化的“就诊后摘要（After-Visit Summary）”，清晰列出用药指导与复诊预警指征，供患者离院后随时查阅。 - **精准等候时间预测**：借助底层的空间临床智能，向候诊患者提供远比传统系统更加精确的实时等候预估，大幅减少服务台的咨询压力，缓解候诊焦虑。 2. **智能语音呼叫铃（Smart Call Light）**： - 完全复用现有的房间硬件，无需额外采购新设备。 - 彻底改变传统呼叫铃“只有一个无上下文的按钮”的现状。患者只需直接呼唤：“Hey Samio，我需要协助下床”、“Hey Samio，我需要护士”、“Hey Samio，我需要止痛药”或“我感觉头晕”。 - 系统能自动对患者诉求进行语义分类，附带上下文精准推送到对应岗位的医护人员手中。这对于肢体无力、无法动手按压传统物理呼叫按钮的重症患者尤为关键。 ## Notable quotes > [00:04](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=4s) 在一位医生看完了 50 位患者的疲惫状态下，或是一名护士经历了长达 12 小时的筋疲力竭轮班后，凭记忆手写录入的数据到底有多可靠？答案是“非常不可靠”。 — Mohamad Elzaim > [00:04](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=4s) 不再开发又一个仅仅帮忙写病历的工具，而是让整个“诊疗环境”本身来完成记录，在临床信息产生的第一时间于物理空间中直接捕捉。 — Mohamad Elzaim > [05:23](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=323s) 医护人员的精力应当 100% 聚焦在面前的患者身上，而不是去“伺候技术”（Babysitting the technology）。 — Mohamad Elzaim > [08:21](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=501s) 医生（MD）的核心价值在于做出临床决策（Making Decisions）。 — Mohamad Elzaim > [10:37](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=637s) 减少收集病史的时间，将更多时间用于做出临床决策，并向患者解释沟通这些决定 — Mohamad Elzaim > [12:18](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=738s) Samio 在为既有医疗行为提供有力抗辩的同时，也揭示了真实的运营管理漏洞。 — Mohamad Elzaim > [16:24](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=984s) 有史以来第一次，管理者能够按岗位角色精确统计患者在一个轮班周期内获得的实际床旁护理时长。 — Mohamad Elzaim > [18:23](https://www.youtube.com/watch?v=rVeXKuK2IHQ&t=1103s) Hey Samio，我需要协助下床 — Mohamad Elzaim ## Mentioned Samio Systems, Mohamad Elzaim, Ahmed Abousaid, Jonathan Grimm, EMR (电子病历系统), Samio 锚点设备 (Anchor Device), 被动式临床工牌 (Passive Clinical Badge), Samio Setup App, LiDAR, RTLS (实时定位系统), Code Blue (代码蓝), Skilled Nursing Facility (SNF) ## Video description Mohamad Elzaim - Co-Founder, Samio Systems What if clinical documentation and workflow intelligence happened automatically, without any reliance on provider behavior? In today’s healthcare environment, healthcare personnel are overwhelmed by administrative burden, with nearly half of their time consumed by documentation, creating inefficiencies and contributing to burnout. In this talk, I will introduce SAMIO Systems and explain how embedding spatial intelligence directly into the clinical environment—through a hardware-enabled, room-level system—fundamentally shifts how data is captured and workflows are executed. I will illustrate how this approach enables autonomous documentation, real-time operational awareness, and system-wide standardization that software-only solutions cannot achieve. Attendees will gain insight into how this infrastructure creates scalable “forced efficiency” and unlocks a new layer of clinical and operational intelligence. If you’re looking to understand where healthcare is heading beyond traditional AI tools, this session will offer a compelling look at what comes next. [Source](https://www.youtube.com/watch?v=rVeXKuK2IHQ)

Source quotes

[0:04] Mohamad Elzaim: 在一位医生看完了 50 位患者的疲惫状态下，或是一名护士经历了长达 12 小时的筋疲力竭轮班后，凭记忆手写录入的数据到底有多可靠？答案是“非常不可靠”。 [0:04] Mohamad Elzaim: 不再开发又一个仅仅帮忙写病历的工具，而是让整个“诊疗环境”本身来完成记录，在临床信息产生的第一时间于物理空间中直接捕捉。 [5:23] Mohamad Elzaim: 医护人员的精力应当 100% 聚焦在面前的患者身上，而不是去“伺候技术”（Babysitting the technology）。 [8:21] Mohamad Elzaim: 医生（MD）的核心价值在于做出临床决策（Making Decisions）。 [10:37] Mohamad Elzaim: 减少收集病史的时间，将更多时间用于做出临床决策，并向患者解释沟通这些决定 [12:18] Mohamad Elzaim: Samio 在为既有医疗行为提供有力抗辩的同时，也揭示了真实的运营管理漏洞。 [16:24] Mohamad Elzaim: 有史以来第一次，管理者能够按岗位角色精确统计患者在一个轮班周期内获得的实际床旁护理时长。 [18:23] Mohamad Elzaim: Hey Samio，我需要协助下床

Source quotes

Dr. Mohammad Elzaim: What if clinical documentation and workflow intelligence happened automatically, without any reliance on provider behavior? [0:04] Dr. Mohammad Elzaim: spatial clinical intelligence [0:04] Dr. Mohammad Elzaim: rather than creating another tool to write the chart, let the clinical environment itself capture and document care in real time. [5:23] Dr. Mohammad Elzaim: babysitting technology [5:23] Dr. Mohammad Elzaim: walking into the room starts the encounter, and walking out stops it. [8:21] Dr. Mohammad Elzaim: a clinical visit does not begin when the physician enters; it begins during rooming, when a medical assistant (MA) or nurse conducts intake. [15:02] Dr. Mohammad Elzaim: cuts both ways [18:22] Hey Samio, I need help getting up [18:22] Hey Samio, I need my pain medications [18:22] Hey Samio, I feel dizzy

Original transcript

[0:04] So, hello everyone. My name is Dr. Mohammad Elzaim. I'm a primary care physician based out [0:08] of Dallas, Texas. Uh I'm the CEO and co-founder of Samuel Systems. I also want to introduce our [0:14] co-founder and CEO Dr. Ahmed Abousaid. He's sitting in the back. [0:18] And our lead engineer, we have Jonathan Grimm also in the back, okay? Uh I'm very excited to [0:23] be here today to share what we are building here at Samuel, but also I'm equally excited [0:27] to be introducing a new way of thinking about health care AI, a concept that we call spatial [0:32] clinical intelligence, where we are embedding technology directly into the clinical environment. [0:39] So, it's not controversial to say that the EMR sits at the heart of every health care system. [0:45] Today's AI tools summarize charts, extract and reorganize clinical data, and automate [0:50] administrative work. And all of this is built on top of what? You guessed it. It's the EMR. [0:56] So, let's take a step back and ask, where does that record from the EMR itself come from? [1:02] Uh usually, it's from a clinician, a nurse, an aide, some health care professional, right? And [1:07] let's take an even further step back. Where does that health care professional get that record? [1:12] And it's not a trick question, I promise. You guys want to throw anything out there? [1:18] All right, it's okay. The answer I'm looking for is is memory. Uh [1:22] physicians usually see a cluster of patients, they write notes, or they might write them at the end [1:26] of the day. That's what I mean when I say memory. Uh so, let's take the human memory for a minute. [1:31] Uh most of our existing AI models in health care are trained on data gathered from the EMR, [1:36] which was produced from the memory of a medical professional. Uh now, here's the [1:40] uh here's the key question. How reliable is a human memory writing charts at the end of a [1:45] 50-patient day? Or a nurse at the end of a grueling 12-hour shift? The answer is [1:51] not very. Yet, our most sophisticated AI tools in medicine are running on data [1:55] entered by hand from memory by stressed work stressed busy and overworked medical professional. [2:02] Yes, AI scribes have helped but recent surveys put day-to-day adoption anywhere between 25 to 40%. [2:08] Uh and they come with their own deficiencies which will come to later. So at Samio [2:14] we thought instead of building one more tool that writes the chart, why not let the [2:18] environment write it? Why not create a system that captures clinically relevant information [2:23] at the very point that information presents itself. This is the fundamental principle [2:28] of spatial clinical intelligence which is the idea that an intelligence space [2:33] or an intelligent clinical environment should understand the physical space, what that space is, [2:37] where care is occurring in that space, who's participating in that care and for how long. [2:42] You'll soon realize that building spatial intelligence in the clinical environment opens up [2:47] uh litany of valuable use cases that we haven't seen before. So how are we building spatial [2:53] intelligence into the clinical environment? So Samio starts with two core components. [2:58] First is the Samio anchor device and you see that here in the figure. These are these little blue uh [3:04] uh squares on the walls. Uh the Samio device is installed in every exam room or patient room. [3:10] It serves as that room's central intelligence hub. The second component is a passive clinical [3:15] badge that's worn by physicians, nurses, medical assistants and any other member of the care team. [3:20] It simply clips onto the badge holder the staff already wears. Users don't have to interact with [3:25] the badge in any way except charging it when the battery indicator is low. It is otherwise [3:29] completely passive. So we'll get to how it works in a minute but let me touch on the installation [3:34] process. So we designed Samio to be fully uh self-installable. A clinical facility receives the [3:42] Samio devices based on the number of rooms they require then a handy person mounts the devices on [3:47] the wall like a picture frame and connects it to a power outlet. It can be installed in any wall. [3:52] It's only restriction it does have to be in a fixed position. So now comes the cool part. [3:57] Our Samuel setup app uses Apple's built-in lidar. Is the video not working? Oh, we know. [4:06] Okay, so the video is not working. But the idea is uh there's a lidar sensor [4:11] uh that sets up the two-dimensional blueprint and you get something like [4:17] like this. Let me go back. Yeah, how do I go back? Is it the red button? [4:26] Oh, okay. Okay. [4:30] All right, there we go. So using our own setup app and an uh iPad or iPhone, [4:34] you can get the two-dimensional blueprint directly from the scan. [4:39] And the cool thing is is then that generates a two-dimensional blueprint of the clinical space, [4:43] okay? And then when you combine it with uh uh the badge and the the anchor device, [4:50] you get something called uh a real-time location tracking system, which is shown here. Um [5:00] Now some of you may be thinking I did this whole run-up and hyped this all up about [5:04] a technology that already exists, which is RTLS. It's true that these systems do exist, [5:08] but they're hard to set up. They have several components, they're expensive, [5:11] but most importantly, they're not designed for the integration that enables a wide set of new [5:17] and powerful use cases. The first of which we'll talk about is transcription. [5:23] So the AI uh the AI scribe market has taken off. The most popular right now are software-based [5:28] solutions and they all share a very common workflow. The clinician has to bring in [5:31] electronic device, usually a laptop or phone, into the room. They may have to enter the patient name, [5:36] visit type, or encounter details. They have to get consent for the AI transcription. They [5:40] may adjust the phone or laptop to make sure it's positioned right. They have to start recording. [5:44] When the visit's over, they have to stop the recording. And I'll admit none of these steps are [5:49] difficult on their own, but whether it's an app, a laptop, or a tool built directly into the EMR, [5:54] the fatal flaw is the same in that these tools require manual steps, and more importantly, they [5:59] demand a constant slice of the user's attention. Imagine a busy ER, clinic, or urgent care, a [6:05] physician seeing a ton of patients a day, getting pulled out of one room, rushed into another, [6:10] stopped with a quick question in the hallway. Now, imagine that physician forgets to hit start [6:14] recording, or forgets to end it, or clicks on the wrong patient, or forgets to pause it when they [6:18] step out and double back, or discuss the case with a nurse while another patient's encounter [6:22] still capturing. That's where the errors happen, and every one of these is a small human slip, [6:27] and every one of them compromises the transcript. So, that, I'll reiterate, is the fatal flaw. [6:31] Unless the user dedicates a significant slice of their attention to managing the technology, [6:36] mistakes will and often do happen. The focus of physicians and any other member of the healthcare [6:42] team should be on the patient in front of them, not on babysitting the technology. So, at Sermo, [6:50] we automated that entire workflow. All anyone has to do is walk in. There's no extra devices, [6:55] there's no buttons, and role-specific medical summaries are generated. [6:59] Because now the environment is intelligent, the physical presence is the trigger for starting [7:03] and ending the recording. There's no buttons on an app or clicks on the screen. The result is a [7:08] fully automated audio capture system for medical transcription we believe is the first of its kind. [7:13] So, from an adoption perspective, uh this is enormous. This reduces a lot [7:18] of the friction associated with adopting new technologies. Imagine telling a 78-year-old uh [7:22] Dr. Jones at whatever hospital has been working there since paper charts to fiddle with an app to [7:27] try to make him more efficient. No. Instead, you tell him, "Here, Dr. Jones, let me see your badge [7:31] clip for a second." Then he goes to the chart and he sees his notes are automatically there. [7:38] So, this table shows uh AI Scribe adoption pattern at an average clinic. I've mentioned before the AI [7:43] scribe adoption rate is anywhere between 25 to 40%, which is represented by the left figure. [7:49] Uh the left side the of the table. Uh the biggest issue is that these AI scribing companies are [7:54] marketed towards individual physicians. So if a facility's gains on efficiency are capped [7:59] by however many individuals say yes or use the system. The right table uh is the same clinic, [8:07] but now you have facility-wide adoption because all you have to do is give them a badge. So the [8:12] facility can get the per uh the efficiency gains uh simply by giving the badge to their physicians. [8:21] All right. So uh quick show of hands. How many of you have been to the doctor [8:26] and felt like you repeated your story a frustrating amount of times? Right? [8:33] All right, here's why that uh that happens. So you may think your visit starts whenever you see [8:37] your doctor, but it actually starts when you're being roomed. When the medical assistant or nurse [8:41] begins intake, and that intake often contains the first freshest version of the story. Whether [8:46] it's symptom severity, medication changes, symptom timeline, allergy screenings, concerns the patient [8:51] may never repeat the same way again. A may uh patient might tell the medical assistant all this [8:57] that's on here on the on the slide, but guess what makes it into the EMR? Just two or three words. [9:05] Uh so that's data loss, and once that information is gone, me the physician have to recreate the [9:10] entire encounter from scratch, which is why you end up telling your story a thousand times. [9:15] So with Samio, every part of the visit is captured. Before I even open the door, [9:20] I have a semi-detailed history of what the patient told my staff. When I walk in, I can review it, [9:25] confirm it, expand on it, or correct it based on my own evaluation. And in the figure you hear [9:29] see here, you see MAs going in Oh, man, it's not playing. Is there any way you can put uh [9:38] Is there any way you can make that figure play? Yeah, it should start automatically. [9:44] All right. Yeah, and the figures are showing uh the MAs going into the room and then the [9:49] they have their own badge and the the device is picking up the transcript. [9:53] It'll be ready for the physician to review. So, uh and then another thing [9:59] uh here's where it gets more powerful. AI just doesn't just capture the history, [10:03] it can make it better. As my MA is rooming the patient, AI systems can guide the conversation [10:08] suggesting the right follow-up questions based on the chief complaint. So, the history that reaches [10:12] me, you can also press play for that, too. I guess we have to do that for every single one. [10:17] So, the history that reaches me is more complete than a standard intake. This means I'm not walking [10:21] in to re-ask the same questions the patient uh had already answered. The the redundant work is done [10:27] before the visit even starts. So, in fact, I would have enough information to place preliminary [10:32] orders before I even enter the room. So, somebody told me that my MD stands for making decisions, [10:37] and I think that's what Samio allows physicians to do more of. Less time [10:41] gathering history, more time making decisions, and communicating those decisions to our patients. [10:46] And then speaking of patients this is another huge benefit of Samio. Now, physicians can [10:51] focus their attention on what matters most uh on people, not on screens and keyboards. [10:56] Uh okay, so that's transcription. Uh I'm going to transition to talk about my favorite feature uh [11:03] we've developed at Samio, and it's real-time translation. Uh so, most clinish clinics and [11:08] hospitals serve multilingual populations, but think about how translation actually works today. [11:12] There's usually one interpreter iPad per floor, one. Uh you have to find it, wheel it in, [11:17] dial into the service, wait uh for the interpreter to connect, and when they and then they have to [11:21] say their 5-minute spiel, and then finally the visit starts. The whole setup can take longer [11:26] than the encounter itself. And guess what happens? Staff start rationing translation. It gets used [11:31] only in the encounters I feel important enough, whether it's a new diagnosis, surgical consent, [11:35] doctor visits. This creates clinical risk. With Samio, that entire barrier disappears. [11:41] That same device that captures the visit for transcription supports real-time multilingual [11:46] communication in every room, all the time. There's no cart to find, no service to dial, and no setup. [11:54] And then here is another figure. How did that one play? And the other one was there. Anyway. [11:58] So, uh this is the figure. You can just read through it. But basically, [12:03] the device has a speaker on it. Uh the physician speaks to the patient in their native language. [12:07] The device spits out in the patient's language. The patient responds in their language, and it [12:11] spits it out in the doctor's language. All right. So, now let's get to some fun stuff. All right. [12:18] So, uh I'm going to show you how Samio would function in a real scenario. So, this is based [12:22] on a real case. So, JM is a 67-year-old man admitted to the room 101 with pneumonia. [12:28] Initially on a small amount of oxygen, but over the next 4 hours, his lungs slowly start to fail, [12:33] and he goes into something called cardiopulmonary arrest. A code blue is called. [12:38] And for those unfamiliar with what a code blue is, if you don't mind playing the video. [12:47] A code blue means a patient's heart or breathing has stopped. It's the emergency of emergencies [12:51] in hospitals. This is one what uh what one looks like. And to be clear, this is just a simulation. [12:56] That's just a mannequin. So, nobody get panicked, okay? All right. [13:03] So, now here's the same code blue through the eyes of Samio. [13:10] So, pay attention to room 101. It's the top left. [13:16] So, you'll see the staff converging on room 101 the moment the code is called. [13:20] Again, spatial intelligence, we know who was present, when they got there, and for how long. [13:26] But Samio doesn't just show us the event. It lets us look back to see what happened. So, [13:31] we can see the hours leading up to the event, every room entry, every bedside evaluation, [13:36] uh laid out as objective room level data. Um Okay, there we go. [13:43] You can see here as they're entering the room, [13:45] it's keeping a log of who was there, how long they were there. Um [13:52] Now, in this case, uh the patient's family filed a lawsuit. The complaint alleges the [13:56] patient was left unmonitored for hours before his arrest. Now, watch what the Samio timeline shows. [14:02] So, we don't have cameras in health care for obvious reasons, but after a bad outcome, [14:06] the question that everyone asks is what actually happened. And that goes for the patient, patient's [14:10] family, the hospital, the doctors, the nurses. Usually, we look at notes and the timestamps [14:16] from the EMR and try to piece it together. But now, we have objective room level data. [14:21] And look at what is showed in this case. There's nine badge-verified room entries, roughly 80 [14:26] minutes of bedside presence over 4 hours, nursing, physician, and respiratory therapy all involved. [14:31] The claim that no one was checking in on him is objectively refuted. The patient was not [14:36] abandoned. And you see here, it was refuted on the top left. Uh but the timeline also showed showed [14:42] something else. There was a 47-minute gap during shift change where the patient was not checked on. [14:47] And that's the key point. Samio defends the care that was given while surfacing the [14:51] real operational issues. The shift change gap becomes a quality improvement finding [14:56] the hospital can actually fix, all while keeping the record honest. So, before Samio, [15:02] I served as uh medical director at two skilled nursing facilities. And at one of them, [15:07] I swear to God, we had a patient who filed some sort of uh abuse allegation every other month. [15:12] Uh every single time the state had to come out and investigate, as they should. [15:16] But each investigation means staff interviews, records pulled, days of administrator burden. [15:21] And all of it is reconstructed from memory and paper. So here we have a complaint. [15:27] All right. So a nursing home resident claims a man named Michael came into my room [15:32] in the middle of the night and hit me. Now watch what the Samuel Look Back shows. [15:37] The room level timeline lays out exactly who entered the room, when, how often, and how long. [15:43] The allegation gets answered with objective data instead of recollection. The investigation that [15:48] used to take days can get resolved in minutes. But here's the part I love because it cuts both ways. [15:53] Incidentally, on that same look back, we can see the CNA isn't rounding on the patient the [15:57] way she's supposed to. The policy in the facility says Q2 rounding or rounding every 2 hours. We see [16:03] here she's rounding every 4 hours. So it surfaced a real gap in care we needed to fix. That's that [16:08] honest record again. It protects you when the claim is exaggerated and it holds you accountable [16:13] when the care falls short. Both of those make the facility better. So everything I showed you so far [16:19] serves the encounter, but there's another layer which is our third product, which is analytics. At [16:25] the basic level that means metrics we simply never had. How much time did the patients spend with [16:30] the provider? How long did nursing spend on the bedside? How long did patients wait in their room? [16:35] What's our real room utilization? But the deeper analytic value is in the workflow analytics. Where [16:41] are the actual bottleneck? Is it provider or is it the rooming? Are rooms sitting idle while patients [16:46] wait? Where do care gaps actually cluster? These are questions every clinic and hospital ask and [16:52] until now the answers have been guesswork. We can spend the whole session on analytics alone, [16:57] but in the interest of time let me show you just two examples. The first is care burden. [17:03] It's another animation you have to press play on. And this is a big deal. So for the first [17:08] time ever we can measure the actual time of care a patient receives across an entire shift, [17:13] every bedside minute per role. And that opens up real applications on the staffing side. We can see [17:18] when one nurse is carrying a disproportionate share of the workload, and we can balance [17:22] assignment based on the actual care burden, not just patient counts. On the clinical side, [17:27] we can chart a single patient's care burden over time. If their care needs are steadily climbing, [17:32] that's a signal worth asking about. Why is this patient requiring more and more attention? [17:37] Sometimes that trend is the earliest sign that something is changing. So, the second is [17:44] path burden. So, how staff actually move through the facility. So, you can watch this nurse. [17:49] Uh she's making trip after trip to the supply room. It's in the bottom right corner. [17:54] So, all shift long. So, that's time walking uh and not spending time with the patients. [18:00] So, we can optimize for that. So, first we can see the pattern, which is that red line. [18:05] Uh it's showing that that's time wasted going to the supply room, and we can do quality improvement [18:09] to improve that. So, if we move the supply room maybe a little bit closer, we can cut that [18:15] uh traveling time the nurse has to do and give her more minutes to actually care for patients. [18:22] All right. And we're wrapping up here, but I'll touch on two more products that we're [18:26] kind of working on. This is just going to be some highlights. Uh so, we're working on a patient [18:30] facing facing app that has two main functions. One is patient education, after visit summary [18:36] generated from the actual encounter, not a generic handout delivered in the patient's preferred [18:41] language with medication guidance, return precautions that they can revisit at home. And [18:46] another cool feature is real wait times. Because Samio provides spatial clinical intelligence, [18:50] patients can see live wait time estimates much more accurately than other time systems. So, [18:56] there's fewer front desk questions, less waiting room frustration, and more patient satisfaction. [19:01] So, another feature we're building is uh a smart call light. Uh again, with the same Samio [19:07] devices. There's no extra devices. So, instead of patient pressing a button with no context, [19:12] the patient can say simply say, "Hey Samio, I need help getting up. Hey Samio, I need my nurse. [19:17] Hey Samio, I need my pain medications or I feel dizzy." So, Samio classifies the request, [19:22] routes it to the appropriate staff member, and provides more context than a traditional [19:27] call light ever could. And this is huge for patients that are not able to physically [19:31] press a call button. Uh so, voice activation would be huge. So, uh [19:36] this about wraps up our presentation. So, uh I hope I communicated the concept clearly. [19:42] Uh I don't think we have any more room for questions. So, if you have any questions, [19:46] just feel free to visit our booth or you can talk to me afterwards. Thank you very much.