Exo

The probe stands without its models, and the AI features are a growing layer on it. Exo Iris uses a piezoelectric micromachined transducer on silicon, and the device images without any model attached. On top of it Exo sells AI suites: Cardiac AI (automated ejection fraction, chamber size, strain and wall thickness), Lung AI (B lines, effusions, consolidations and lung pathology in real time), Fluid Management AI (IVC collapsibility, bladder volume and vascular access) and SweepAI, which gives instant feedback as the user sweeps.

Several detect findings rather than only measuring structures, and SweepAI takes part in acquisition itself, which makes the layer substantial. Exo Works, the workflow product, is vendor neutral software for documentation, billing, quality assurance and education, and is not a model.

Advisory by design, with a comparative claim that pulls against the oversight the design assumes. The AI features surface on the device in real time beside the image they come from, and the clinician decides; nothing acts on a patient. SweepAI gives instant feedback as the user sweeps, turning an inadequate image from a silent failure into visible feedback, which is a real oversight control.

The company states that its pleural effusion and consolidation AI showed superiority to clinicians alone. A clinician told a tool is better than they are has little reason to override it, and the cases where overriding matters are the unusual ones where a model is least reliable; the intended users include nurses and emergency medical technicians, who may not be able to second guess a finding on the image.

What is not published is behavior at the limits: what a feature does with an inadequate or out of scope image, whether it declines rather than guesses, and what confidence information reaches the user.

The AI features are named one by one, and nothing is published about what produces them. The home page names the Cardiac AI Suite (automated ejection fraction, chamber size, strain and wall thickness), the Lung AI Suite (B lines, effusions, consolidations and lung pathology), Fluid Management AI (IVC collapsibility, bladder volume and vascular access) and SweepAI, which gives feedback during the sweep, and the transducer is described specifically as a piezoelectric micromachined design on silicon.

No model architecture, training data, validation population, version history or update practice is published there, and features arrive on devices already in the field. The 510(k) summaries for the AI features could describe the training and validation data.

The AI features are the company's own cleared models, and the parties around them are not named. Exo states that SweepAI and its cardiac and lung AI are FDA cleared on the Exo Iris, and describes its June 2025 pleural effusion and consolidation AI as the first such clearance for point of care ultrasound. Models the company has cleared as its own devices cannot be swapped without a regulatory consequence.

What the home page does not publish is the rest of the chain: the cloud provider behind Exo Works, a subprocessor list, where the training data for each feature came from, and whether any model component was licensed. Exo Works is also vendor neutral, taking studies from other manufacturers' probes through interfaces that are not named. A subprocessor list could sit in the Security Trust Center linked in the site footer.

The company's strongest evidence statement is a comparative one, and its figures are not on the home page. Announcing its June 2025 clearance for detecting pleural effusion and consolidation or atelectasis, which it describes as the first for point of care ultrasound AI, Exo stated that the tool showed superiority to clinicians alone.

That is a claim about a study, and the study's comparator, case mix, reader experience and effect size decide what it means: an emergency physician and a trainee are different baselines. The company's announcement of that clearance and the 510(k) summary behind it are where the design and figures would be stated.

The other evidence is operational: Wellstar MCG Health, formerly Augusta University Health, implemented Exo Works across its health system, a named deployment of the workflow product. No outcome study showing changed diagnosis, treatment or patient outcome, and no peer reviewed publication, is on the home page.

Strong security governance is reported around the data, and nothing is published about what the models may learn from it. The custodial half rests on audited certifications for Exo Works and a Security Trust Center. The stewardship half is not addressed on the home page: whether customer images are used to train, tune or evaluate the company's models, whether any such use stays inside a customer boundary, how long studies are retained, and whether a customer may decline.

The question carries weight here because the company keeps adding cleared models, and because Exo Works is vendor neutral, accumulating studies acquired on other manufacturers' probes as well as its own; a customer scanning on another maker's hardware may not realize the question applies to them. Safety engineering, meaning guardrails and how a safety event is handled, is not published. The privacy policy linked in the site footer may state a training use position.

The security program is audited, and a business associate position is not stated on the home page. Exo Works holds HITRUST risk based two year (r2) certification and a SOC 2 Type 2 report, announced in 2023. HITRUST maps onto HIPAA's security and breach rules, but it is a security assessment rather than a statement of the company's role for customer data.

The company states plainly that the customer is responsible for securing the phones and tablets the probe pairs with, a useful boundary for a product acquiring clinical images on personal hardware. The privacy policy and the Security Trust Center linked in the site footer are where business associate status and an agreement would be set out.

Three audited credentials are named, each released on request. The company's Security Trust Center lists ISO/IEC 27001:2013, a SOC 2 Type II report and HITRUST certification, and each is obtained through a support request rather than opened directly; the company announced in 2023 that Exo Works earned HITRUST r2 certification and completed its SOC 2 Type 2 report. The ISO/IEC 27001 entry names the 2013 edition, whose certificates had to move to the 2022 edition by October 2025, and the entry does not show whether the certificate has moved.

No validity dates, scope statements, penetration testing or vulnerability disclosure policy are published, and nothing says whether the Exo Iris and its AI features fall inside the certified scope. The company names the customer as responsible for securing the phones and tablets the probe pairs with.

The device clearances are identifiable, and the AI clearances are stated without their numbers. The Exo Iris was cleared in August 2021 (K211527), with a further clearance in 2022 (K222198) adding imaging modes and indications. The company's AI page states that its cardiac AI (ejection fraction and stroke volume), its lung AI (A lines, B lines, pleural effusion, and consolidation or atelectasis) and SweepAI are FDA cleared and work on the Exo Iris, and the company describes its June 2025 clearance for pleural effusion and consolidation or atelectasis as the first for point of care ultrasound AI, which is its own claim.

No clearance number, version or intended use is given for any AI feature, so published material does not tie a feature to its 510(k).

Testing is described and its results by group are not. The company's AI page says validation included diverse demographics (gender, ethnicity, location), body mass index and health conditions including heart failure with reduced ejection fraction, coronary artery disease and cardiomyopathy, and that the models were developed on more than 100,000 images from point of care users in emergency and critical care units. It reports no performance for any of those groups and states no conclusion that performance held across them. Separately, the company describes SweepAI's acquisition feedback as reducing operator dependence, which concerns variation between operators rather than patients.

The gap matters because ultrasound image quality varies with body habitus, and these detection features inform admission, antibiotic and diuretic decisions, often in the hands of nurses and emergency medical technicians without an expert reader, so a model that misses findings more often in one group produces silent undertreatment there.

Nothing published says what happens when an AI feature is wrong, and these features make findings rather than only measurements. The Lung AI Suite detects B lines, effusions and consolidations in real time, and the Cardiac AI Suite reports ejection fraction, chamber size, strain and wall thickness. A missed effusion or consolidation, or a falsely reassuring ejection fraction, is a diagnostic failure that can shape admission, antibiotic or diuretic decisions.

No warranty covering model output, accuracy commitment, indemnity or route to dispute a result is published on the home page. The company states that its pleural effusion and consolidation AI showed superiority to clinicians alone; marketed to nurses and pre hospital staff who may have no expert reader beside them, a comparative claim of that kind makes the allocation of responsibility more pressing, not less. A published error rate with its denominator may sit in the 510(k) summaries for the AI features, whose clearance numbers the home page does not give.

Exo Works is vendor neutral, and the systems it connects to are not named on the home page. The workflow product handles documentation, billing, quality assurance and education for point of care ultrasound programs, and is described as integrating with nearly any point of care ultrasound device and with common hospital record and imaging archive systems, reachable from a phone, tablet or browser.

Working with other manufacturers' probes addresses the real problem of mixed probe estates, where exams never reach the record consistently, and Wellstar MCG Health implemented it across its health system. No record system, archive vendor, interface standard or third party device list is named on the home page. The Exo Works page in the site navigation is where integrations would be listed.

Continuity is described, and where data rests is not on the home page. Exo states that it maintains a framework to limit the impact of disruptive events on its operations and that its business continuity plans are validated through its third party audit certifications.

Exo Works is reachable from a phone, tablet or browser with studies stored in the cloud, and the probe pairs with a clinician's own device; the company places responsibility for securing those phones and tablets with the customer. No infrastructure provider, region, residency position, tenancy or recovery objective is given on the home page, which matters for a company selling internationally and into rural and under resourced settings. The Security Trust Center linked in the site footer is where hosting and residency would be documented.

A device price is published, and the recurring structure is not. Exo Iris launched in September 2023 at a published starting price of $3,500, and the site footer links an online store that sells the probe directly. What is not published on the home page is everything recurring: whether the AI suites come with the device, are unlocked by subscription or are licensed per feature; whether a membership or subscription is needed and what lapsing it does; and how Exo Works, the enterprise workflow product deployed across Wellstar MCG Health, is priced.

Coverage is organized around who holds the probe, with regulatory validation behind the AI features. The stated users are physicians, nurses, emergency medical technicians and other clinicians in emergency, acute care, outpatient and home settings, with rural and under resourced settings called out. Clinical coverage centers on heart and lung, the pair that matters for undifferentiated breathlessness, with fluid management (IVC and bladder volume) and vascular access alongside.

The AI features carry FDA clearances; home and pre hospital use are named as targets without published deployments, and Exo Works extends to other manufacturers' probes without a published device list.