Three years after an initial forecast of AI's impact on healthcare, a new assessment finds that adoption has surged in clinical documentation, prior authorization, and patient engagement-but the technology is largely accelerating broken processes rather than repairing them. The review, which revisited the original 2022 predictions, highlights a market where AI triage companies grew 20x year-over-year to a $100 million-plus category, while prior authorization AI spending exploded tenfold to $100 million in 2025.
Patient care and delivery
AI-driven triaging and access have become a $100 million-plus market. Platforms like Counsel Health, Doctronic, and Torch Health assess symptoms conversationally and route patients to appropriate care. Scheduling automation tools such as Assort Health and Hello Patient eliminate manual appointment booking, addressing the same no-show and slot-matching problems that earlier startups tried to solve.
Patient engagement has seen more fundamental change. AI care navigation platforms-Hippocratic AI, Ferry Health, Aidify-handle the gaps between clinical encounters. They call with results, schedule follow-ups, and coordinate care transitions. Physicians now routinely convert discharge summaries into patient-friendly explanations using tools like DoximityGPT, creating warm handoff letters for primary care providers in under a minute. As these tools handle tasks like triage and follow-up, the need for healthcare staff to understand AI's role grows. AI for Healthcare Courses can help clinicians and administrators navigate this shift.
Yet most of these tools are efficiency AI. They make broken processes faster but do not fix the dysfunction. As the analysis put it: "We're automating abandonment instead of eliminating it."
Research, diagnostics, and treatment
The diagnostic and clinical decision support space evolved toward AI-augmented medical references rather than replacing clinical reasoning. UpToDate launched Expert AI, which emulates expert clinicians' reasoning with full transparency into assumptions and sources. OpenEvidence raised $210 million for rapid literature synthesis. Doximity acquired Pathway Medical for $63 million and integrated it into DoxGPT. ClinicalKey, DynaMed, AMBOSS, and Glass Health all launched AI-powered clinical decision support tools.
The transparency of these tools-grounding answers in trusted content and showing their reasoning-is critical for adoption. When physicians and trainees use them to learn and improve their knowledge base, the next generation of clinicians could be among the brightest ever.
Clinical and non-clinical workflows
AI scribes are now ubiquitous. Abridge, Suki, Nabla, Doximity, and dozens of others automate clinical documentation. The tools save physicians time, but they do not question why clinicians spend half their time on notes in the first place.
The prior authorization AI arms race has escalated dramatically. Spending in this area grew from $10 million in 2024 to $100 million in 2025. Companies like Latent Health, Tandem, and Mandolin auto-fill payer forms by pulling data from EHRs. Payers respond with their own AI, such as Optum Real, to deliver instant coverage validation. Both sides are investing billions to fight the same battle more efficiently without asking whether the battle should exist.
CMS recently brought AI-driven prior authorization to Traditional Medicare through the WISeR Model, paying AI contractors a percentage of denied claims to review services "prone to unnecessary use." This imports a criticized Medicare Advantage feature into a system that was already working well.
"AI won't fix healthcare by making broken processes faster. It will only fix healthcare if it exposes the misaligned incentives we've been able to blame on complexity," the review concluded.
Why this matters for healthcare professionals
The three-year checkup shows that AI is stripping away the excuse of complexity. When a tool can automate prior authorizations in seconds and slash administrative costs, the remaining inefficiency is no longer a technical problem-it's a problem of incentives. Healthcare organizations and the professionals within them must use AI not as a patch for broken workflows, but as a reason to question whether those workflows should exist at all. The ones that confront misaligned incentives head-on and realign them around efficiency and patient care will outperform those that simply run their broken systems faster.
Your membership also unlocks:
