Editorial brief · Methodology

How we evaluate AI tools for clinical practice

MedAI Verdict is a curated reference. We do not run every tool in our own clinical practice. Here is what we do instead, what we weight, and what we explicitly do not claim.

Last updated May 18, 2026

01 / Sources

The sources we draw on

Each review draws from a stack of named public sources. Weights below reflect how much each source family contributes to a tool’s position in our comparisons.

Source 01 Peer-reviewed literature Weight 25%: PubMed-indexed studies that evaluate the tool in clinical settings. JAMA, NEJM, BMJ, specialty journals when relevant.
Source 02 Reddit clinician communities Weight 20%: Public clinician and trainee discussions where AI tools are evaluated in working clinical context. Sentiment extracted, quotes attributed.
Source 03 Public review aggregators Weight 15%: G2, Capterra, Software Advice, TrustRadius profile pages. Star-ratings and recent qualitative reviews.
Source 04 Vendor stability signals Weight 15%: Funding rounds, leadership changes, acquisitions, EHR-marketplace certifications (Epic App Orchard, Cerner Code).
Source 05 Physician networks Weight 10%: Sermo and Doximity, closed physician-only platforms. Quoted with attribution and anonymization.
Source 06 Vendor documentation Weight 5%: Pricing pages, security attestations (HIPAA, SOC 2, HITRUST, FDA-510(k)), EHR-integration disclosures. Used as the factual source for specs and pricing, not as evidence for ranking position.
Source 07 Clinician YouTube reviews Weight 5%: Long-form clinician demos and product walkthroughs from board-certified or trainee-credentialed channels.
Source 08 Specialty society guidance Weight 5%: AAFP, AAP, ACP, AMA, ACR, ACS recommendations or endorsements when published.

02 / Principles

Six editorial principles that govern every review

Principle 01

Aggregation over self-testing

We do not run AI tools in our own clinical practice for the purpose of these reviews. Sites that claim hands-on testing of every tool in their index typically rely on the same public sources we do, with less transparency about it. We name our sources and link to them.

Principle 02

MD editorial sign-off

Every published review is reviewed by Henrik R., MD, a board-certified physician (Switzerland), before publication. The MD-verified badge appears only on tools whose reviews have completed sign-off within the past six months. Identity verifiable to vendor partners and editorial inquiries on request.

Principle 03

Affiliate transparency

Some outbound links are affiliate links. We disclose this inline at the point-of-occurrence (a sponsored badge next to each link) and at the bottom of every page. Rankings are editorial and never sold. We decline sponsorships from tools that fail our published evaluation criteria.

Principle 04

Source attribution per claim

Quantitative claims (pricing, integration counts, certification status) link to the source. Qualitative quotes from clinician reviews carry the source forum, subreddit, or platform with a date. We do not republish entire reviews; we quote with attribution under fair-use commentary.

Principle 05

Living documents, not frozen reports

Every review carries a last-verified timestamp per data category. Our scrapers re-run vendor documentation and community sentiment on a monthly cadence. Reviews older than 180 days carry a stale-data warning until refreshed.

Principle 06

Transparent uncertainty

Every tool page includes a What we have not verified block: data points the public sources do not let us validate (private SLA terms, enterprise pricing tiers below NDA, real-world latency in specific EHRs). We name what we do not know.

03 / Transparency

What we do not claim

01We do not claim "hands-on testing" of every tool in our index. When a tool review includes first-hand observations from an MD on our editorial team, the review is marked with that note.
02We do not claim our scoring is objective. Weights are editorial judgments published openly. Reasonable clinicians can disagree.
03We do not claim our reviews replace formal evaluation by your IT, compliance, or legal teams. Use this site to narrow a shortlist, not to make a final purchasing decision.
04We do not claim AI-tool reviews are medical advice. Reviews are about commercial software, not clinical guidance.
05We do not claim independence from commercial relationships. Affiliate links exist and we disclose them. We claim that our rankings are not for sale.

04 / Affiliate disclosure

Where money may change hands

Some outbound links to AI-tool vendors are affiliate links. If you sign up through one of these links, we may receive a referral commission. The price you pay is identical. The following commercial relationships exist:

Direct affiliate or referral agreements with selected vendors. These links carry a sponsored badge.
Sponsorship slots on category landing pages (clearly marked, category-exclusive maximum 1 per quarter).
Newsletter sponsorships in our forthcoming clinician newsletter (subject to the same editorial firewall).

Editorial rankings are not affected by these commercial relationships. We decline sponsorships from tools that fail our evaluation criteria. If you spot a vendor relationship we have not disclosed, please email editorial@medaiverdict.com.

05 / Corrections

Corrections policy

Pricing, integration, and certification data change rapidly. If you are a vendor, clinician, or reader and you spot an inaccuracy, we correct it within seven business days of verification. Corrections are logged at the bottom of each affected review with the original wording struck through.

Material errors that affect a tool’s ranking position trigger a re-publication notice, not a silent edit.

06 / Contact

Reach the editorial team

Editorial: editorial@medaiverdict.com
Corrections: corrections@medaiverdict.com
Vendor relations: partnerships@medaiverdict.com
Press: press@medaiverdict.com

Editorial services

Custom AI evaluation, on retainer.

Beyond the public reference, our editorial team takes a limited number of private evaluation engagements each quarter for practices, health systems, and vendors. MD-led, methodology-bound, independent.

Consulting fees fund the public editorial operation. We decline engagements that would compromise editorial independence.

Start a conversation

Tier 01
Single-tool evaluation
Independent deep-eval of one AI tool for your practice or specialty. Spec audit, EHR-integration risk assessment, pricing analysis, deployment timeline, and a written recommendation memo signed by our editorial team.
For: Solo practitioners and small groups
Tier 02
Procurement review
Comparative eval across 3 to 5 shortlisted vendors. Vendor reference calls, RFP scoring framework, security and compliance gap analysis, deployment-risk matrix, executive summary for leadership sign-off.
For: Health systems and large group practices
Tier 03
Vendor MD sign-off review
For vendors: independent MD review of clinical claims, security attestation, integration depth, and editorial cross-check against our methodology. Output published as a verified entry with disclosure note.
For: Healthcare AI vendors