Rad AI

Rad AI is a healthcare technology company that develops generative artificial intelligence software for radiology departments, hospitals, and diagnostic imaging practices. Founded in 2018, the company focuses on improving the way radiologists create reports, document imaging findings, and manage patient follow-up recommendations. Its technology is designed to reduce repetitive administrative work, improve reporting consistency, and help radiologists spend more time interpreting medical images. Rather than replacing radiologists or making independent clinical decisions, Rad AI provides software that assists medical professionals throughout the radiology reporting and follow-up process.

The company was co-founded by Dr. Jeff Chang, a practising radiologist, and Doktor Gurson, a technology entrepreneur. The idea emerged from the challenges Dr. Chang experienced while working in radiology, particularly increasing imaging volumes, repetitive documentation, and physician burnout. The founders recognised that generative AI and machine learning could help automate parts of the reporting process while allowing radiologists to retain control over clinical interpretation and final reports. Rad AI was subsequently developed around specialised AI models trained for radiology and healthcare language rather than general-purpose text generation.

Rad AI’s product portfolio is organised around three major areas: Rad AI Reporting, Rad AI Impressions, and Rad AI Continuity. Each application addresses a different part of the radiology workflow. Reporting supports the creation and completion of medical imaging reports, Impressions generates draft report summaries from dictated findings, and Continuity identifies and manages recommendations for additional imaging or clinical follow-up. These products can be used separately or as complementary components of a wider radiology workflow.

Rad AI Reporting is the company’s comprehensive reporting platform. It is designed to help radiologists create complete medical imaging reports through natural dictation, speech recognition, structured templates, and generative AI. The software supports both free-form dictation and structured reporting, allowing radiologists to work in a style that suits their preferences and established clinical processes. It can convert spoken findings into organised report content and generate impressions while preserving the radiologist’s preferred language, terminology, and reporting style.

One of the platform’s important capabilities is its personalised reporting experience. Rad AI uses historical reports and dictation patterns to adapt generated content to individual radiologists. This helps reduce the need to rewrite generic AI-generated text and allows the software to produce reports that more closely reflect the user’s established approach. Practices can also migrate existing templates, macros, and picklists or configure new ones. This is useful for radiology organisations that need consistency across their reporting operations without forcing every radiologist to abandon familiar documentation methods.

Rad AI Reporting also includes speech intelligence designed specifically for medical imaging terminology. Radiologists can dictate findings naturally while the system processes speech in real time. Its speech technology is tuned for radiology language and specialist terminology, helping reduce interruptions and the editing associated with conventional dictation systems. The platform also supports voice commands and multi-author workflows, which can be valuable in departments where reports require contributions from more than one professional.

Another capability is the conversion of free-form dictation into structured reports. Instead of requiring radiologists to dictate information in a rigid sequence, the software can organise dictated content into the appropriate sections of a report. It also supports automated calculations and formulas, helping reduce repetitive manual work. For follow-up examinations, the system can carry forward unchanged findings so that radiologists can concentrate on the differences between the current examination and earlier studies.

The reporting platform includes real-time quality checks intended to identify inconsistencies and omissions before a report is finalised. These checks can include patient information, report structure, and clinical guideline recommendations. The system can also insert relevant consensus-based recommendations into reports where appropriate. These functions are intended to improve consistency and support quality assurance, but the radiologist remains responsible for reviewing the content and approving the final report.

Rad AI Impressions is a specialised application that automatically generates the impression section of a radiology report from the findings dictated by a radiologist. The impression is an important part of a radiology report because it summarises the relevant findings and communicates the principal conclusions to referring healthcare professionals. Creating this section can require considerable concentration, especially when reports involve complex findings or multiple recommendations. Rad AI Impressions is designed to reduce this repetitive writing task by generating a draft impression within seconds.

The application works alongside existing radiology reporting software. As a radiologist dictates findings, Rad AI Impressions processes the report text and creates a proposed impression that can be inserted into the existing report editor. The radiologist can then review, edit, and approve the generated content. This approach allows imaging departments to introduce AI-assisted impression generation without necessarily replacing their existing reporting platform or voice recognition system.

Personalisation is also an important part of Rad AI Impressions. The software can use de-identified historical reports to learn the writing patterns and preferences of individual radiologists. This helps it produce impressions that reflect the user’s usual terminology and style rather than relying exclusively on standardised wording. Organisations can also configure the system around their reporting practices and clinical preferences. The application supports several imaging modalities, including computed radiography, CT, PET/CT, MRI, ultrasound, and mammography.

Another feature of Rad AI Impressions is its ability to insert consensus-based clinical guideline recommendations into the impression when relevant. This can help radiologists include appropriate follow-up guidance and make reports more consistent. The application also provides feedback and monitoring capabilities that allow organisations to evaluate adoption, efficiency, and reporting consistency. Its purpose is to assist with documentation, not independently determine a diagnosis or replace the radiologist’s professional judgement.

Rad AI Continuity addresses a different challenge: ensuring that important follow-up recommendations in radiology reports are acted upon. Medical imaging examinations sometimes reveal incidental findings that require additional tests, specialist consultations, or repeat imaging. These recommendations can be difficult to manage when healthcare organisations rely on manual tracking, separate spreadsheets, or disconnected communication processes. If a recommendation is not followed, patients may miss an important opportunity for further assessment.

Rad AI Continuity uses artificial intelligence and natural language processing to identify follow-up recommendations in radiology reports. It extracts relevant information, classifies findings, and organises them into worklists that can be reviewed by nurse navigators and other authorised healthcare personnel. The system supports more than 50 categories and over 250 subcategories of findings, allowing organisations to manage different follow-up programmes through a centralised platform.

The application provides customisable navigator worklists that help staff prioritise patients, assign tasks, review recommendations, and monitor outstanding cases. It also supports multi-dimensional searches across patient information, providers, and pathologies. These features help healthcare teams organise large numbers of follow-up recommendations and identify cases that require attention. The system can track multiple findings for the same patient independently, which is useful when one imaging examination produces several separate recommendations.

Communication and follow-up tracking are central to Rad AI Continuity. The platform can support automated communication with patients and referring providers through configured electronic health record messages, printed letters, and other supported channels. Healthcare teams can review and approve scheduled communications where required. The system also maintains activity histories and audit trails, giving organisations visibility into how recommendations are managed and whether the required follow-up has been completed.

Rad AI Continuity receives status updates through healthcare system integrations, including HL7 events. These updates can help keep follow-up records aligned with scheduling and results workflows. Its bi-directional integration capabilities allow information to move between the platform and existing electronic health record or radiology information systems. This reduces the need for staff to repeatedly enter the same information into separate systems.

The platform also includes specialised support for lung cancer screening follow-up. Its lung cancer screening module is designed to identify eligible patients, organise screening workflows, monitor follow-up requirements, and support communication with patients and providers. This can help healthcare organisations manage screening programmes more consistently and monitor whether patients complete recommended examinations.

Rad AI Continuity is designed to support a closed-loop follow-up process. It identifies a recommendation, places it into an appropriate worklist, supports communication and coordination, and tracks the case until it is resolved. This can help reduce missed follow-ups, improve patient communication, and provide administrators with better visibility into programme performance. Clinical decisions and the responsibility for patient care remain with qualified healthcare professionals.

Rad AI’s products are designed to operate within existing healthcare technology environments. The company supports integration with PACS, RIS, and EHR systems and uses established healthcare interoperability standards, including DICOM, HL7, and FHIR. Rad AI Reporting also supports standards and technologies such as DICOM Structured Reporting, SAML, and OpenID Connect. These integration options are important for larger hospitals and multi-site radiology practices that need to connect new software with existing clinical applications, identity systems, and reporting infrastructure.

The company also supports integrations with established radiology reporting applications. Rad AI Impressions can work with platforms such as Nuance PowerScribe and Fluency through a lightweight desktop client. This allows radiologists to continue using familiar reporting environments while receiving AI-generated impressions. Rad AI’s partnership with Siemens Healthineers has also expanded the availability of its solutions through a strategic reseller arrangement. The company has additionally announced a partnership with RSNA Ventures to bring radiology knowledge and educational resources into its reporting workflow.

Security and data protection are important considerations because Rad AI processes sensitive healthcare information. The company states that its platform is designed to support HIPAA requirements and uses enterprise security controls and encryption practices. Its product information also identifies SOC 2 Type II certification. Healthcare organisations should still review the applicable security documentation, contractual arrangements, access controls, and deployment requirements before connecting the software to patient information systems.

Rad AI is primarily intended for radiologists, radiology practices, hospitals, imaging networks, healthcare executives, and clinical operations teams. Radiologists can use its reporting and impression-generation applications to reduce documentation tasks, while administrators and patient navigators can use Continuity to organise follow-up programmes. IT teams can evaluate its cloud-native architecture and standards-based integrations when planning deployment across multiple facilities.

The company reports adoption across more than 200 health organisations and says it works with over 40% of U.S. health systems and nine of the ten largest U.S. radiology practices. These figures indicate a substantial enterprise presence, although actual availability and adoption may vary by product and healthcare organisation. Rad AI has also expanded its technology through product development, partnerships, and investment in specialised speech recognition and generative AI capabilities.

Rad AI is not a general-purpose chatbot or a standalone medical diagnostic system. Its products are specialised workflow tools designed for the documentation and coordination activities surrounding medical imaging. The generated content and extracted recommendations must be used within appropriate clinical review processes. Radiologists remain responsible for interpreting images, reviewing AI-generated reports, making clinical decisions, and signing final documentation. Patient navigators and other healthcare professionals remain responsible for their respective follow-up and communication duties.

Pricing

Rad AI uses an enterprise software model for hospitals, radiology practices, and healthcare organisations. Its website does not publish fixed subscription prices or a standard self-service purchasing plan. Prospective customers are directed to request a demonstration and contact the company to discuss their requirements.

The commercial cost may depend on the selected products, number of radiologists, imaging volume, facilities, integrations, deployment requirements, and support arrangements. Organisations considering the platform should request a customised quotation and evaluate implementation and ongoing operating costs.

Review

Rad AI is a specialised radiology workflow platform with a clear focus on reducing the administrative burden associated with medical imaging. Its combination of reporting, impression generation, and follow-up management gives it broader workflow coverage than a tool focused only on speech-to-text transcription. The three product areas address different operational needs, making the platform relevant to both radiologists and healthcare administrators.

Its strongest advantage is the way it fits into established radiology environments. Support for existing reporting systems, structured templates, voice recognition platforms, and healthcare interoperability standards can make adoption more practical for organisations that do not want to replace their entire technology infrastructure. Personalised report generation is also valuable because it helps maintain individual reporting preferences rather than forcing radiologists to work with generic templates.

Rad AI Continuity adds another important capability by connecting imaging recommendations with patient follow-up activities. Its worklists, communication tools, status tracking, and audit trails can help healthcare organisations manage processes that might otherwise depend on manual coordination. This makes the platform relevant not only to radiology productivity but also to operational oversight and continuity of care.

There are limitations to consider. Rad AI is designed for enterprise healthcare environments, so smaller practices may find the implementation and purchasing process less accessible than that of a self-service application. Pricing is not publicly disclosed, making it difficult to compare costs without contacting the company. Deployment also requires technical coordination with existing reporting, EHR, RIS, and PACS systems.

Another important consideration is clinical oversight. AI-generated impressions and extracted follow-up recommendations require appropriate review and verification. The technology can support reporting and coordination, but it should not be treated as a replacement for professional interpretation or clinical responsibility. Organisations must also assess security, interoperability, and performance within their own workflows before deployment.

Rad AI is a strong option for healthcare organisations seeking to modernise radiology reporting and follow-up operations. Its specialised models, personalised reporting capabilities, and enterprise integrations make it particularly relevant to hospitals and large imaging practices. Its suitability for a particular organisation will depend on its reporting infrastructure, clinical requirements, integration needs, and available budget.