The healthcare AI market makes more sense when companies are grouped by the job they are hired to do. This map organizes 27 companies into seven buyer categories: ambient documentation, imaging and diagnostics, clinical decision support, patient access, revenue cycle, life sciences, and remote monitoring. It is a working procurement map, not a ranking and not an endorsement.
That distinction matters because a market map built around the word AI quickly becomes useless. A company drafting notes, a company detecting findings on CT images, and a company automating payer calls may all use similar model language. They solve different problems, involve different owners, and carry different evidence and regulatory burdens.
1. Ambient Clinical Documentation
The ambient clinical documentation category includes Abridge, Microsoft Nuance, Ambience Healthcare, Suki, Nabla, and DeepScribe. These systems listen to an encounter and draft a note for clinician review. The business case is usually clinician time, retention, and documentation burden. The buying questions are note quality, correction time, specialty fit, patient consent, audio retention, model training, EHR integration, coding effects, and whether review remains genuinely human rather than ceremonial.
2. Medical Imaging and Diagnostic AI
The medical imaging and diagnostic AI category includes Aidoc, Viz.ai, PathAI, Lunit, Annalise.ai, and RapidAI. Products in this group may triage cases, flag findings, support interpretation, or coordinate time-sensitive care. Buyers need the exact FDA-authorized use when authorization is required, not a general claim that the company is cleared. Local prevalence, scanner mix, patient population, false alert work, and integration with PACS or pathology systems can change the practical result.
3. Clinical Decision Support and Predictive Risk
The clinical decision support category includes Epic and Bayesian Health in this edition. The small count does not mean the market is small. It reflects how often predictive models arrive inside the EHR or are built by a health system rather than purchased as a visible point solution. Prospective validation, alert burden, subgroup performance, clinical ownership, and written stop conditions matter more than a retrospective accuracy number.
4. Patient Access and Engagement
The patient access and engagement category includes Notable Health, Hyro, Luma Health, and Artera. The category spans intake, scheduling, referrals, messaging, and call-center automation. A successful interaction is not a conversation completed by a bot. It is a patient reaching the correct service with less friction and a reliable option to reach a person when the system is wrong or the need is urgent.
5. Revenue Cycle and Administrative AI
The revenue cycle and administrative AI category includes AKASA and Infinitus Systems. These companies automate coding, claim review, benefit verification, authorization, and payer communication. Volume is a weak success metric. The financial case should include denial rate, rework, exceptions, audit findings, days in accounts receivable, net collections, and the human effort still required around the automation.
6. Drug Discovery and Life Sciences AI
The drug discovery and life sciences AI category includes Tempus, Recursion, Isomorphic Labs, Insilico Medicine, and BenevolentAI. This group mixes data platforms, research partnerships, technology companies, and drug developers with internally owned pipelines. Any comparison should separate AI-originated programs from AI-assisted research, platform revenue from pipeline value, and early discovery milestones from clinical endpoints.
7. Remote Monitoring and Virtual Care
The remote monitoring and virtual care category includes Biofourmis and Cadence. The algorithm is only one component. Devices, patient adherence, home connectivity, escalation rules, clinical staffing, reimbursement, and responsibility for after-hours alerts determine whether a monitoring program creates care or creates another inbox.
The Four Filters That Travel Across Every Category
- Evidence: Was the product evaluated in a population, workflow, and setting comparable to ours, and were outcomes measured beyond model accuracy?
- Regulatory and accountability: What rules apply to the exact function, who owns the decision, and what can pause the system after go-live?
- Data and contracts: What data enter the system, where are they processed, how long are they retained, and can they be used for model improvement?
- Workflow and economics: What human work remains, what new work appears, and which clinical or financial result changes after all exceptions are counted?
How This Map Connects to AI Medicine Now
AI Healthcare Now owns the buyer and market layer. AI Medicine Now owns the clinical vendor layer. Eleven companies already have clinical profiles there: Abridge, Aidoc, BenevolentAI, Insilico Medicine, Notable, Nuance DAX, PathAI, Recursion, Suki, Tempus, and Viz.ai. Their names link to those profiles rather than creating a second clinical page here. The remaining companies have concise buyer profiles on this site.
Method and Update Standard
Companies are assigned to the category representing the primary job described in this directory, even when a product spans several functions. Inclusion is based on category relevance and a working official company source. It does not imply a commercial relationship, product recommendation, comparative performance finding, or complete coverage of the market. Category details and links were verified on September 12, 2026.