Agentic AI for the full physician-development decade

From Year One of Medical School
to Year One of Independent Practice.

Turning the least supported decade of a physician's career into the most supported one.

Capture the moment Perform the work Preserve judgment

Medicus Tiro builds Agentic AI for the physician-development decade—medical school, applications, residency, fellowship, and the transition into independent practice. Our products are designed to perform that decade's administrative work while preserving the competence, judgment, and trust required to train great physicians.

Long-term vision: one decade, one connected journey

A physician-owned portfolio designed to carry trusted information across every transition—medical school, the Match, residency, fellowship, credentialing, and the move into practice. That record does not span the full journey today; it is what the architecture is intended to enable.

Company and product

Medicus Tiro is the company, supporting the physician across the full decade. GME Manager is its first product and current commercial entry point, focused on residency.

Explore the Physician Journey See Our First Product
Now GME Manager — our first product
  • Built Operating foundation
  • 6 programs Family Medicine pilot initiated
  • $500K HealthStream strategic investment
  • Live hStream integration
  • Run alongside Incumbent system stays in place
Full traction detail →

These points describe GME Manager, Medicus Tiro’s first product, in residency — not the wider planned portfolio. Status as of August 2026.

The Decade

From the first year of medical school to the first year of independent practice

Ten years. Six stages. Multiple institutions, systems, and sets of responsibilities. Our long-term vision is infrastructure for a connected journey—a physician-owned portfolio designed to carry validated evidence across every stage.

Every transition forces the physician and the institution to begin again.

Records are recreated. Evidence is left behind. Credentials are collected and reverified again and again. Medicus Tiro is building toward continuity across the decade.

🎓

Medical School

Years 1–4: coursework, clerkships, away rotations, USMLE, and the first clinical evidence

The decade starts here
📋

Applicant

MSPE, ERAS, and the Match—residency and fellowship applications with verified credentials

🏥

Residency

Training management, milestones, evaluations, procedures

🩺

Fellowship

When applicable: subspecialty training, advanced procedures, research

📑

Transition to Practice

Credentialing, licensure, board certification, privileging, and the employment transition

🚀

First Year of Independent Practice

Early-career support, clinical documentation, mentorship, and enduring alumni relationships

One decade. Repeated transitions.

Each is a handoff between institutions, systems, and responsibilities—and a point where evidence, context, and support are currently lost.

  • Medical education
  • Clinical rotations
  • Residency applications
  • Residency
  • Fellowship
  • Credentialing
  • Licensure
  • Board certification
  • Employment transition
  • Early independent practice
The Problem

One decade. Repeated transitions. No continuous support system.

Burnout is the symptom. Twenty-five years of accumulated administrative burden is the cause. That burden accumulates at every stage of the physician-development decade, and falls on both the physicians progressing through it and the people responsible for supporting them.

Medical students enter training with mental health equal to or better than their age-matched peers. By the end of the first year that advantage is gone, and on the measures tracked through training it does not return to baseline. The burden continues through applications, residency, fellowship, credentialing, and the transition into practice. They deserve better.

The burden changes form at every stage, but it never stops accumulating.

The physician moves through five or six institutions, systems, and sets of responsibilities in ten years. At every handoff the record is rebuilt, the support structure resets, and the administrative work is added to rather than absorbed.

  • Stage 1

    Medical School

    Students arrive with mental health equal to or better than their peers; by the end of Year 1 that advantage is gone. 27.2% screen positive for depression, and only 16% of them seek treatment. Clerkship and away-rotation evidence scatters across every site the student touches.

  • Stage 2

    Applications & the Match

    The MSPE is assembled by hand from records held in separate silos. The applicant re-supplies information the institutions already hold, and none of the clinical evidence built over two years follows them into residency.

  • Stage 3

    Residency

    The most documented stage: 47–50% burnout, a multi-institutional median of 217 faculty evaluations per trainee per year, and CBME raising documented observations from roughly 350 to 1,750+ per program per year.

  • Stage 4

    Fellow

    Subspecialty training adds another application cycle and another institution. The same core record is re-entered for the fellowship application, and the procedural and evaluation evidence built across residency does not follow the trainee into the new program.

  • Stage 5

    Transition to Practice

    The same core record is re-entered again for state licensure, board certification, credentialing, and privileging. A physician completes roughly six credentialing applications a year, and leaving training triggers a credentialing crunch that routinely runs three to six months.

  • Stage 6

    First Year of Independent Practice

    Transition shock: the workload of a twenty-year veteran, with the overnight loss of program directors, documentation buffers, and clinical mentorship. 30–40% of new physicians leave their first job within five years.

Every figure in this strip is documented, with its research base, in the evidence behind the least-supported decade below.

Inside residency, four roles carry different parts of the same load

Residency is where the accumulation has been measured most closely, and where it is visible on every side of the room: trainees, faculty, coordinators, and program directors each carry a different part of the same structural burden.

Resident
47–50%
Burnout—exceeds practicing physicians
80-hour weeks, $223K debt on $67K salary, retrospective evaluations, no support cliff at PGY-2.
Faculty
34%
Of new faculty leave within 3 years
217 evaluations/trainee/year (multi-institutional median), from memory, on legacy tools. 70–75% of time on clinical RVUs; education unprotected.
Coordinator
36%
Screen positive for burnout; 28% likely to leave within a year (Slavin et al., 2025)
Complex program management under clerical job titles. Fragmented tech stack, no career path, no training pipeline. 30%+ annual turnover (ACGME ADS, 2021–22).
Program Director
11–14%
Annual turnover; 4.5-year median tenure
Expanding mandate on shrinking protected time. Designed as a career—has become a rotation.

Burnout is what gets measured. Administrative burden is what we built this company to remove.

Explore the evidence behind the least-supported decade.
The Evidence

A population that arrives well, and measurably deteriorates during training—documented across seven dimensions.

Single-study claims are named inline where they appear; the named sources and the numbered sources 1–4 are listed at the foot of this page.

Everyone is talking about physician burnout. What's far less understood is how administrative burden compounds across the full physician-development decade, and how acutely it is felt in residency, on every side of the room. The infrastructure supporting those transitions has not kept pace with the work now required.

We call it the least supported decade: from the first year of medical school through the first year of independent practice, with too few tools built for the physicians moving through it or the people supporting them.

Everyone knows physician burnout is a crisis. Here's what almost no one is talking about: resident burnout is worse — 47–50%, against 43% among practicing physicians. And the most frequently reported reason so few of those suffering seek help is fear that treatment will harm residency prospects or licensure. The measured distress is not evenly distributed across a career. It concentrates in this decade.
47–50%
Resident & fellow
burnout rate
27%
Medical students screen
positive for depression
16%
Of depressed students
who seek treatment
3×
Suicide risk vs.
age-matched peers
01
They Enter Healthy—Health Deteriorates During Training
Mental health peaks at matriculation, bottoms out by Year 1, and never returns to baseline.
Matriculating medical students begin with mental health equal to or better than age-matched peers. Physical, emotional, and overall health peak at matriculation, bottom out by the end of Year 1, and do not return to baseline at any point during training. In a national comparison, medical students reported the highest depression and suicidal ideation, residents the highest burnout and fatigue, and early-career physicians the lowest fatigue — a pattern in which the training decade is the peak-distress window, with measured distress lower after it ends.
02
Burnout Peaks Across the Full Decade
47–50% of residents vs. 43% of practicing physicians. The training years are worse—not better.
Medical student burnout: 37–50%. Resident and fellow burnout: 47–50%. Practicing physicians: 43%. It is not linear—burnout spikes to 40–45% during PGY-2 and PGY-3, a period when clinical responsibility scales up and structured supervision scales down. Residents score worse than attendings on depersonalization (40.8% vs. 28.6%) and work-life satisfaction (33.6% vs. 42.2%, against 58% for the general U.S. workforce), at 60–80 hrs/week, with little autonomy and little pay, for years.
03
Depression & Suicide: 2–3× the General Population
27.2% of medical students screen positive for depression. 300–400 physicians die by suicide annually.
Across 195 studies and 129,000+ medical students, 27.2% screen positive for depression and 11.1% report suicidal ideation—rates 2–3× higher than demographically comparable peers, and medical students are 3× more likely to die by suicide than age-matched controls. Among residents, suicide is a leading cause of death. Across the profession, 300–400 physicians die by suicide annually—an entire medical school class every year. These rates improve after training.
04
The Financial Insolvency Gap
$223,000+ in debt on a $67,000 salary—the most extreme ratio of any profession, sustained for 3–7 years.
Average medical education debt now exceeds $223,000, serviced on first-year resident salaries of ~$67,000. This creates the most extreme debt-to-income ratio of any profession at any career stage, sustained for 3–7 years with no meaningful capacity to pay down principal. Debt above $250K is independently associated with 1.24–1.47× higher burnout odds. In 2025, 75% of residents cited student loan debt as a top financial stressor, up from 49% the prior year. The financial pressure eases markedly once independent practice salaries arrive—but by then it has run for a decade.
05
The Help-Seeking Penalty
Only 16% of depressed medical students seek treatment—and fear of career consequences is the most reported barrier.
Across 33 studies, the top barrier is fear that mental health care will harm residency prospects or licensure. Where state licensing boards asked about lifetime history of any diagnosis or treatment, physicians were measurably less likely to seek care. Nearly 40% of physicians reported unwillingness to seek psychiatric treatment due to licensure fears. Few professional decades pair this level of need with this level of disincentive to seek support.
06
The Effects Don't End When Training Does
30–40% of new physicians leave their first job within five years. Burned-out physicians are ~1.5× more likely to leave or change practices.
Resident burnout is directly associated with higher medical-error rates, increased attrition, and career regret—14% of PGY-2 residents overall, and up to 33% in some specialties, report they would not choose medicine again. Burnout also predicts departure: in a 2026 JAMA Internal Medicine study of 19,929 family physicians, 43.5% reported burnout, and burned-out physicians were nearly 1.5× more likely to leave or change practices—70% more likely to exit medicine entirely and 40% more likely to move to a new practice. New attendings face “transition shock”: the same high-stakes workload as 20-year veterans, but overnight loss of program directors, documentation buffers, and clinical mentorship. Mid- and late-career physicians consistently report lower burnout and higher satisfaction—consistent with a crisis concentrated in this decade rather than distributed across a career.
07
CBME Is About to Multiply the Documentation Load
217 faculty evaluations per trainee per year is the multi-institutional median today—and CBME pushes documented observations from ~350 to 1,750+ per program per year.
The administrative burden isn't holding steady—it's accelerating. The multi-institutional median is already 217 faculty evaluations per trainee per year, reconstructed from memory on legacy tools, and CBME replaces periodic evaluation with continuous evidence: for a typical 21-resident Family Medicine program, documented observations rise from roughly 350 to 1,750+ per year, a 5× increase. Core-outcome attestation begins in Family Medicine in June 2026, with procedural-competence attestation required beginning in June 2027. Each of the 14 ABFM core outcomes needs continuous evidence. The mandate improves training — and adds to the documentation burden carried by the people in this decade.

The least supported decade doesn't only affect physicians. It reaches every patient they see, every system they work in, and every dollar invested in producing them.
That's the reality. Here's what we're building to change it.

See the product vision we're building toward

Figures above are drawn from peer-reviewed research, national surveys, and federal data sources. Where a single study carries a claim it is named inline — Slavin et al., 2025 (coordinator burnout); ACGME ADS, 2021–22 (coordinator turnover); JAMA Internal Medicine, 2026 (family-physician burnout and attrition); ABFM (June 2026 and June 2027 attestation requirements). The numbered sources 1–4 used elsewhere on this page are listed in full under Sources. Complete citations for the remaining figures are available on request.

Our Product Vision

Purpose-built for every stage

Our long-term vision is a portfolio of AI-native products supporting physicians from medical school through the transition into independent practice—growing from the foundation GME Manager established, and connected over time by the Validated Lifelong Portfolio.

Scroll the map sideways to see all five physician stages →

Product concepts
Medical Student Rotation Manager Planned · Partner-dependent
GME Recruiting Manager Planned · Partner-dependent
GME Manager Current flagship · Built · Pilot initiated

The product Medicus Tiro brought to market, and the foundation the rest of this map is intended to grow from.

Alumni Network Planned · Partner-dependent
Physician stages
Student Medical school
Applicant ERAS & the Match
Resident Residency
Fellow Subspecialty
New Physician Independent practice
Additional product concepts
Medical School AI Success Suite Long-term vision
GME Onboarding Manager Planned
GME Schedule Manager Planned
GME Finance Manager Long-term vision
GME Reimbursement Manager Planned
Resident Outboarding & Career Launch Long-term vision
Shared layers — intended to span every stage
Vision: Courseware, Assessments, Exam Prep & CME

A learning and assessment layer intended to run beneath every stage. Not built today; partner-dependent.

Vision: The Validated Lifelong Portfolio

Our long-term vision for a physician-owned record spanning applications, milestones, credentialing, privileging, enrollment, and career transitions.

  • Current flagship — built, pilot initiated
  • Planned — designed, partner-dependent, build follows commitment
  • Long-term vision — intended future architecture
Medicus Tiro’s product vision mapped across the physician journey. GME Manager is the current flagship; every other concept here is planned or long-term vision, including the Validated Lifelong Portfolio — connective architecture we intend to build, not a record that operates across the journey today.
  • Decade-wide vision

    One validated, physician-owned record spanning medical school through the first year of independent practice, carried by the physician rather than left behind in an institution's system.

  • Built foundation

    The operating foundation across Layers 1–3 is built, including the shared data layer and the resident-facing physician-owned record inside GME Manager. That record covers residency, not the decade.

  • Current functionality

    What is live in the initiated pilot is a deliberately narrow slice: Procedure Logging and AI-drafted Evaluations, with the capture, confirmation, review and export workflows around them.

  • Planned applications

    Portfolio applications at the other stages of the decade—student portfolio, applicant portfolio, portable credentials, career records—are designed and partner-dependent. Build follows commitment.

That record will only hold if something upstream captures the evidence in the first place. The same architecture is designed to do that at every stage of the decade — today it does it in one.

How It Fits Together

Four Layers. One Physician Journey.

Everything Medicus Tiro builds sits on four layers. The first three are the brand platform made structural—capture the moment, perform the work, preserve judgment—and the fourth carries the result forward. The same layers are intended to apply at every stage: the moment changes from a clerkship observation to a procedure to a credential; the architecture underneath does not.

  • Capture near the moment
  • Convert unstructured activity into structured evidence
  • Perform the administrative work
  • Preserve human judgment
  • Maintain continuity across transitions
  • Support variable levels of AI autonomy
  • Build toward a portable, validated longitudinal record
Layer 1

Capture the Moment

Capture architecture

Clinical and educational evidence is captured where the work happens—usually by voice, usually on a phone—instead of being reconstructed from memory later.

  • MomentReady™ Capture architecture built around the moment itself
  • Private recovery workflow Between capture and submission — Moment Inbox™
  • Planned ambient-documentation workflow ScribeReady — planned
  • What is captured Procedures, faculty observations, evaluations
Layer 2

Perform the Work

Products and agents

Captured information becomes completed administrative and educational work—drafted, structured, and routed to the person who has to decide.

  • GME Manager Our first product — focused on residency
  • Agent Studio Agent-development platform inside GME Manager (residency scope)
  • Specialty-specific agents Workflow capabilities tuned to one specialty
  • Curated data sets The intelligence each agent works from
Layer 3

Preserve Judgment

AI-control framework

AI assistance is calibrated to risk and reversibility, and the program sets the level. Competency decisions stay human.

  • Variable Autonomy™ AI-control framework, three modes per agent
  • Workflow execution and explanation layers The two rails agents execute and explain on — Action Rail™ and Insight Rail™
  • Entrustment Everywhere™ Reusable educational capability for supervision judgment
  • Faculty review AI prepares; faculty confirm and decide
Layer 4

Carry Evidence Forward

Evidence continuity layer — intended future architecture

What the first three layers produce is intended to travel with the physician rather than stay behind in an institution's system. Today it operates inside GME Manager, at the residency stage; extending it across the decade is the vision.

  • Validated Lifelong Portfolio Long-term vision — evidence continuity across the decade
  • What it is designed to carry Accomplishments, assessments, clinical evidence, credentials
  • Physician as the Source of Truth™ Planned — the record is designed to belong to the physician, not the institution

Where the layers stand. The foundation across Layers 1–3 is built, and the initiated Family Medicine pilot inside GME Manager activates a deliberately narrow slice of it: Procedure Logging and AI-drafted Evaluations. The remaining capture workflows, and the portfolio applications at the other stages, are planned.

Trust, Calibrated

AI you control, calibrated to the risk.

The 2026 clinical-AI synthesis puts it directly: how clinicians and AI interact matters as much as what the model can do.2 Variable Autonomy™ is that design.

Three modes per agent, set by the responsible human—in residency, the Program Director. Higher the stakes, tighter the oversight. Lower the stakes, more autonomy. Reversibility decides where the line moves. The same control model is intended wherever a Medicus Tiro product operates, because the requirement it answers—human judgment stays authoritative—does not change by stage.

Verify Mode
AI suggests. You approve.

Milestone and evaluation decisions. Faculty evaluations. Competency attestation. High stakes, low reversibility.

Augment Mode
AI handles routine. Flags exceptions.

Procedure logging. At-risk resident signals. Compliance monitoring. Medium stakes — exceptions escalate to the PD.

Autonomy Mode
AI executes. Escalates only when needed.

Reminders. Routine follow-up. Administrative routing. Low stakes, high reversibility.

Every workflow configurable by task, role, specialty, or program.

How Variable Autonomy works in practice

Frontier models can outperform attendings on controlled reasoning tasks and still degrade safely-deployed performance when the interaction architecture is wrong. The 2026 ARISE State of Clinical AI synthesis says it directly: “How clinicians use AI is as important as what the model can do.”2 Variable Autonomy is the implementation of that finding for medical education.

Verify Mode keeps the human in the reasoning loop on every judgment-building decision. Augment Mode lets AI process the routine while exceptions and discrepancies escalate to the Program Director. Autonomy Mode runs the workflow, with the PD reviewing exceptions.

2 Brodeur PG, Goh E, Tat E, et al. State of Clinical AI 2026, ARISE Network, January 2026. Full source list →

Our Approach

How we build differently

Eight principles that guide every product decision. The AI does the work. You make the calls.

01

AI-Native Architecture

AI is the foundation of every product, designed in from the start rather than bolted onto an existing system.

02

Agentic AI

The AI does the work. You make the calls.

More on Agentic AI

Agentic AI handles the administrative tasks that hold back excellence and efficiency, and Variable Autonomy lets organizations adopt at a pace that works. See AI you control, calibrated to the risk below.

03

Curated Data Sets and Proprietary Benchmarks

Specialty-specific benchmarks and curated data create a differentiated intelligence layer.

More on curated data sets

Specialty-specific benchmarks, historical outcomes, and curated data create an intelligence layer that would require meaningful domain investment to reproduce, and it improves as we add specialties and deepen each one.

04

Physician-Owned Data

Credentials and career documentation are designed to belong to residents and physicians, not institutions.

More on physician-owned data

Today this principle is implemented inside GME Manager, where the resident-facing record belongs to the resident. Extending it into the Validated Lifelong Portfolio—a physician-owned record designed to create 35+ year relationships built on trust and portability—is our long-term vision.

05

Mobile-Native Direction

Products are designed around where physicians, residents, and faculty actually work—mostly away from a desktop.

More on the mobile-native direction

GME Manager's interactive prototype demonstrates the broader mobile direction, with voice input and phone-first workflows; Procedure Logging and Evaluations are in the initiated pilot's scope. Offline capture is planned.

06

Institutional Integration Architecture

We build deep program intelligence, and partner with the platforms already operating at institutional scale.

More on institutional integration

The institutional memory layer for GME won't be built by aggregating CSV uploads into a dashboard. It requires deep, AI-generated program intelligence—competency trajectories, workforce readiness, predictive compliance—flowing into enterprise platforms that already serve thousands of health systems. We build that intelligence, and partner with the platforms that already operate at institutional scale.

07

Designed for Low-Friction Adoption

Products are designed to operate without patient PHI and to reduce implementation burden.

More on low-friction adoption

Products support guided, AI-assisted configuration, and institutional security, privacy, governance, contracting, and IT requirements are confirmed with each organization. Future pilot openings follow a free 30-day model with no credit card and no commitment to convert.

08

Capture the Moment

Capture important observations, decisions, and evidence close to when they occur—at any stage of the decade—so unstructured activity becomes structured evidence that can prepare the work each stage requires. The responsible physician keeps every consequential judgment.

Examples across the physician journey

MomentReady™ is the architectural expression of this principle: capture an important moment once, then prepare the appropriate downstream work for a human to review and decide.

Across the broader vision, this principle can support medical-school rotation evaluations, interview evaluations for residency applications, AI-supported interviewing and interview coaching, resident-centric ambient-scribe tools, Teaching-Physician Billing Capture, Inpatient Rounds & Handoff, and a private capture-recovery workflow—Moment Inbox™.

Status. Inside GME Manager, the initiated Family Medicine pilot applies this to Procedure Logging and AI-drafted Evaluations. The other applications listed above are planned or long-term direction across the decade, not current functionality.

The Portfolio

One architecture. Four stages of the decade.

Building for the whole decade, stage by stage: what exists today, what is designed and awaiting committed partners, and where each piece sits in the physician's journey. Status is labelled on every card.

Stage 1 · The decade begins here

Medical School

And so does the fragmentation: clerkship data sits with one hospital, away-rotation data with another, grades with the school, and none of it follows the student forward. Our medical-school work is designed to begin the physician-owned portfolio at the first clerkship rather than the first day of residency.

  • Clerkship and away-rotation management
  • Student-owned portfolio
  • Clinical evaluations
  • Procedure and experience capture
  • MSPE support
  • Applicant portfolio at ERAS
  • Transition into residency
Partner-dependent · Build follows commitment
🎓

Medical Student Rotation Manager

AI-native rotation strategy for medical schools: competency-first planning, Preceptor Intelligence, and a Match Intelligence Suite with AI-drafted MSPEs, around a student-owned portfolio running from first clerkship through ERAS. Designed to complement clinical placement platforms, not replace them.

Status: designed and researched, not built, contracted, or in pilot. Build follows commitment.

Stage 2 · Where our first product operates

Residency and Fellowship

The opportunities beside GME Manager apply the same architecture to adjacent training and workforce contexts.

  • Procedure logging
  • AI-drafted evaluations
  • Competency evidence
  • CCC support and individualized learning plans
  • Entrustment Everywhere™
  • Agent Studio
  • Resident portfolio
Now · Built · Pilot initiated
GME Manager

Performs administrative work inside the program, starting with Procedure Logging and AI-drafted Evaluations, and runs alongside the residency-management system a program already uses.

See the GME Manager section →
Partner-dependent · Build follows commitment
🩺

APP Residency Manager

Purpose-built Agentic AI for NP/PA postgraduate residencies and fellowships—550+ programs across 35 specialties and 48 states, a segment with limited purpose-built technology. Same agent as GME Manager: new surface, new context, no new architecture.

Partner-dependent · Build follows commitment
❤️‍🔥

Life Support Instructor Hub

A two-sided marketplace that turns residents into credentialed life support instructors (BLS, ACLS, PALS, NALS) and matches them with organizations needing skills evaluators on demand, with duty-hour guardrails throughout. Residents earn. Programs stay compliant.

The planned GME agent roadmap that extends GME Manager appears further down the page, after the product and its current traction.

Stage 3 · The handoff out of training

Transition to Independent Practice

Leaving training triggers the heaviest reconstruction of the decade. The same core record — medical school, residency application, residency, fellowship application, state licensure, board certification — is re-entered at every step, and the move into practice triggers a credentialing crunch that routinely runs three to six months. This is where the Validated Lifelong Portfolio, our long-term vision, is designed to matter most: the evidence should travel with the physician.

  • Credentialing
  • Licensure
  • Board certification
  • Applicant and professional records
  • Physician as the Source of Truth™
  • Employment transition
  • Transfer of validated evidence
Partner-dependent · Build follows commitment
🔐

Physician as the Source of Truth™

Credentials issued once, carried by the physician, and designed to be verified wherever presented. Built on W3C Verifiable Credentials standards, so validated evidence follows the physician across every transition in the decade.

Partner-dependent · Build follows commitment
🎓

Alumni Network

Trusted relationship infrastructure between training programs and their graduates: designed so medical schools, residency programs, and fellowship programs can stay connected with alumni after training ends, support mentorship with current trainees, and follow how graduates' careers develop—without relying on stale contact lists or public social-media searches. Career support is the secondary extension, on permission-based sharing the graduate controls.

Status: designed and researched, not built, contracted, or in pilot. Build follows commitment.

Stage 4 · The decade closes here

First Year of Independent Practice

Support does not follow the physician out of training. New attendings face “transition shock”: the same high-stakes workload as a twenty-year veteran, but an overnight loss of program directors, documentation buffers, and clinical mentorship — and 30–40% of new physicians leave their first job within five years.

  • Early-career support after training
  • Transition shock
  • Clinical documentation
  • Billing and coding capture
  • Oversight and mentorship
  • Continuing professional development
  • Alumni relationships and early-practice outcomes
Partner-dependent · Build follows commitment
🤝

APP Oversight AI

AI-powered chart review designed to turn APP oversight from a regulatory burden into clinical value. Physicians in 38 states must oversee APP charts; structured AI review runs at scale and surfaces only the cases needing attention.

Partner-dependent · Build follows commitment
🎯

Precision Learning Engine

AI-powered precision learning for CME, regulatory training, life support certifications, sim lab, and vendor in-service—designed to identify what you need and support reporting to licensing boards.

Status: designed and researched, not built, contracted, or in pilot. Build follows commitment.

Build follows commitment.

Development advances when a committed pilot or design partner provides the validation, operational context, and financial commitment required to proceed. Apart from GME Manager, nothing on this page is built, contracted, or in pilot today, and these opportunities will be advanced sequentially rather than simultaneously.

That is the company and the decade it is built for. What follows is its first product — and the evidence that the architecture above can be executed.

Now Our First Product · Focused on Residency

AI That Performs the Work of Residency.

Procedure Logging and AI-drafted Evaluations, captured on a phone, alongside the system your program already runs

Why the entry point is residency.

Residency is one high-burden, commercially actionable stage of the decade, and where the shared architecture is being proven. It is not where the decade begins, and not the boundary of what Medicus Tiro is building.

First Product · Residency

GME Manager

Runs alongside the systems you already use: mobile-first capture, specialty-specific intelligence, and AI oversight controlled by the program.

It performs the administrative work of residency, starting with two workflows: Procedure Logging and AI-drafted Evaluations. Every judgment—approval, assessment, entrustment—remains with faculty and the program.

See the broader GME Manager direction
What changes for each role. Residents log a procedure by voice at the bedside instead of reconstructing it days later. Faculty speak an observation and receive a rotation-calibrated draft to rate, edit, and submit rather than starting from an empty form, retaining every rating and approval. Program administrators see what is complete and what still needs action. Program directors receive more complete, source-linked evidence for competency and readiness decisions, with oversight preserved through Variable Autonomy™.

Initial pilot scope: Procedure Logging and AI-drafted Evaluations, with the supporting capture, review, and export workflows around them.

Broader product direction: built for the Competency-Based Medical Education era, the two-rail AI architecture pairs a workflow execution layer (Action Rail) with an explanation layer (Insight Rail), while Variable Autonomy lets program directors control AI independence. The direction includes CCC support, ABFM attestation support, duty-hour intelligence, the resident-facing physician-owned record inside GME Manager—the first step toward the Validated Lifelong Portfolio, which remains a long-term vision—specialty-specific curated data, and additional MomentReady workflows. These are the broader platform direction, not the initial pilot.
Run alongside. Replace when ready.

Programs don't need to replace MedHub, New Innovations, or another incumbent platform to begin, and expand on demonstrated value—full replacement is an eventual option, not a prerequisite.

How information returns to the incumbent system

Completed information can be returned to the incumbent through structured exports or an agreed transfer cadence, and that system can remain the system of record.

Phase 1 Pilot Cohort Fully Allocated

The initial Family Medicine cohort is fully allocated and moving through configuration, baseline measurement, and readiness. Programs interested in the next opening can sign up for the waitlist through GME Manager.

Join the Next Pilot Cohort
6 ACGME Core Competencies 19 FM Subcompetencies / Milestones (ACGME) 14 ABFM Core Outcomes (June 2026; procedural competence deferred to 2027) Two-rail AI: Action Rail™ + Insight Rail™ 13 Live Curated Data Sets 📱 Mobile-Native Direction Procedure Logging · Pilot initiated Evaluations · Pilot initiated No-PHI Design
Explore Agent Studio
Initial Pilot Workflows
Procedure Logging + Evaluations
Built · Pilot initiated
Workflow 1 Procedure Logging
  1. Voice captured at the bedside Captured
  2. Structured log prepared AI
  3. Resident confirms Ready for confirmation
  4. Faculty approval requested Awaiting faculty review
Workflow 2 End-of-Rotation Evaluation
  1. Faculty observation captured by voice Captured
  2. Ratings and rotation context applied Calibrated
  3. AI draft prepared Ready for confirmation
  4. Faculty reviews, edits, and submits Awaiting faculty review

AI prepares the draft. Faculty retains every judgment.

Runs alongside MedHub, New Innovations, or the incumbent system of record. No migration required  ·  No-PHI design  ·  Illustrative
How a program gets started — the four-step adoption path
Adoption path

How a program gets started

Four steps from first conversation to a reviewed pilot. The incumbent residency-management system stays in place throughout.

  1. Readiness conversation

    Confirm program priorities, current systems, workflow pain points, security requirements, and whether the two initial workflows fit. Fit is assessed on both sides; not every interested program will be a match for this cohort.

  2. Configuration and baseline

    Configure the program, roles, and workflow requirements, and agree the baseline measurement approach and readiness checks while the incumbent system remains in place. Institutional security, privacy and IT requirements are confirmed with each organization.

  3. Run alongside

    Launch voice-first Procedure Logging and AI-drafted Evaluations alongside the incumbent system, with residents and faculty retaining confirmation, approval, editing and submission. No migration required, a No-PHI design, and completed information returns through structured exports or an agreed transfer cadence.

  4. Review before expansion

    The program and Medicus Tiro review adoption, completion, timeliness, burden, satisfaction, workflow reliability and operational fit before deciding whether to continue or expand. Expansion follows demonstrated value; full platform replacement remains an option rather than a requirement.

Start narrow. Measure value. Expand only when earned.

What the initiated pilot will measure — six categories, results pending
Measurement framework · Results pending

What the pilot will measure

The initiated Family Medicine pilot tests whether the two initial workflows improve completion, timeliness, burden, satisfaction and data quality while preserving faculty judgment. The categories below are what will be evaluated — not findings.

  • Category 1 · Results pending

    Adoption

    Whether residents and faculty use the workflows consistently enough to evaluate operational value.

  • Category 2 · Results pending

    Procedure logging

    Completion, timeliness, required-field quality, correction rates, and faculty-approval workflow performance.

  • Category 3 · Results pending

    Evaluations

    Completion, timeliness, draft usefulness, faculty editing requirements, and submission reliability.

  • Category 4 · Results pending

    Administrative burden

    Changes in time and follow-up effort for residents, faculty, program administrators and program directors.

  • Category 5 · Results pending

    Satisfaction

    Resident, faculty, administrator and program-lead feedback on usefulness, usability, trust, and willingness to continue.

  • Category 6 · Results pending

    Data and workflow reliability

    Completeness, transcription quality, workflow exceptions, export readiness, and operational support needs.

How this will be evaluated. Success is assessed using baseline and pilot-period comparisons. No outcome results are presented before the pilot produces them.

Built Platform Foundation Broader Agent Direction

GME Manager Agent Studio

Agent-development platform inside Layer 2 and part of GME Manager, so its scope is residency. The foundation is built; the initiated Family Medicine pilot activates a narrow slice of it. Agent Studio is designed to give programs a second entry path: start with a single purpose-built agent and grow into the broader GME Manager experience.

Level 2 · Built platform foundation

One foundation, two intended ways in

The operating foundation across Layers 1–3 is built. It is what both entry paths are designed to run on.

Broader direction
AI That Performs the Work of Residency
GME Manager

Milestone tracking, CCC meetings, ACGME data exports, and duty-hour compliance are designed to expand from the same captured moment the initial pilot workflows produce.

or
Built foundation
Start with One Agent
Agent Studio

The platform itself: specialty-specific configuration, a workflow execution and explanation layer (Action Rail™ and Insight Rail™), an AI-control framework (Variable Autonomy™), and a shared data and workflow architecture — so a program can begin with the single agent that addresses its most urgent problem.

See the Agent Studio roadmap, status labels, and shared foundation
  • Initial pilot Built and in the initiated Family Medicine pilot
  • Built foundation Built platform or operating capability
  • Broader direction Enabled by the platform; not initial pilot scope
  • Planned Designed, not established as operational
  • Partner-dependent Build follows commitment
Level 3 · Planned and partner-dependent agents

Each agent is designed to make every other agent smarter

Every agent reads from and writes to a shared data layer. That's not a bundle. That's a flywheel. Apart from the Evaluation Agent, the agents below are part of the broader GME Manager direction rather than initial pilot scope.

Planned
📋
Attestation Agent
Designed to work from all 14 ABFM core outcomes (procedural-competence item deferred to June 2027) and map resident evidence to attestation requirements, reducing preparation time ahead of the June deadline.
Planned
👥
CCC Agent
Designed to prepare CCC packages with milestone suggestions, narrative summaries, and evidence citations, organizing evidence for review and reducing preparation time.
Initial pilot
📝
Evaluation Agent
AI-drafted evaluations grounded in logged procedures, rotation data, and competency benchmarks. Faculty review and approve—not write from scratch. One of the two workflows in the initiated pilot.
Planned
⏱️
Duty Hour Agent
Designed to apply all 8 ACGME duty-hour rules in real time, with violation detection, predictive alerts, and audit-ready compliance reporting.
Planned
📊
APE Report Agent
Designed to assemble Annual Program Evaluation reports from milestone data, survey results, and program outcomes, reducing manual assembly ahead of board review.
Partner-dependent
🔮
Additional Agents
The architecture is built to grow. Additional agents advance when a committed pilot or design partner provides the validation, operational context, and commitment required to proceed.
Built platform foundation
Shared data layer  ·  Workflow execution — Action Rail  ·  Explanation — Insight Rail  ·  AI-control framework — Variable Autonomy  ·  13 Live Curated Clinical Data Sets  ·  Physician-owned record inside GME Manager — the first step toward the Lifelong Portfolio

The foundation is shared, and every agent is designed to run on it. Agent Studio is intended as another way to reach the same destination.

Advisor Mode: run alongside. Replace when ready.

Many programs are contractually committed to New Innovations or MedHub for years. Advisor Mode is how they get the intelligence without the migration.

Structured Data Exchange. Agents are designed to run alongside an existing platform through structured exports or an agreed transfer cadence, and completed information can be returned to the incumbent, which remains the system of record. Institutional security, privacy, and IT requirements are confirmed with each organization.

Predictive & Generative Intelligence. AI-drafted evaluations, CCC prep packages, ABFM attestation mapping, and compliance prediction—added without disrupting resident workflows.

13 Live Curated Clinical Data Sets. ABFM procedures, ACGME milestones, USPSTF guidelines, G2211/APCM billing codes, and more, plus two additional documentation-coaching sets—a data foundation that would require meaningful domain investment to reproduce.

A Wedge, Not a Forklift. Programs experience AI intelligence with low disruption, and the path to broader GME Manager adoption is open as value is demonstrated. Run alongside. Replace when ready.

Advisor Mode is a deployment option, not a separate product. Every Agent Studio agent is designed to support it.

For investors: three entry paths, one destination

Agent Studio adds a second go-to-market path alongside the full platform sale. Programs ready for comprehensive residency management adopt GME Manager directly; programs that need to start smaller enter through a single agent and expand as each agent compounds the value of the others. Advisor Mode opens a third path: programs committed to incumbent platforms can gain Agent Studio intelligence while that platform remains the system of record. Three entry paths, one destination, one compounding data moat.

For investors: why the ACGME makes the case for us

The ACGME's coordinator FTE tables define a minimum administrative capacity, not a minimum headcount: a program must demonstrate that the work gets done, not that a specific number of people are employed to do it. AI that absorbs the administrative load lets a program meet its required capacity without adding staff — converting an unfunded compliance burden into a productivity gain, and adoption that pays for itself out of capacity the regulator already requires.

Now Current Traction · Our First Product

Where GME Manager stands today.

These figures describe GME Manager in residency, not the full Medicus Tiro portfolio.

Built

Operating foundation developed

6 Programs

Family Medicine pilot initiated July 2026

$500K

Strategic investment from HealthStream

Live

hStream integration operating

Run Alongside

Works with incumbent residency-management systems

No-PHI design  ·  Family Medicine first  ·  Professional judgment stays with faculty

View Investor Overview See the Full-Decade Portfolio

Status as of August 2026

The Evidence

Why now, and why this architecture.

Healthcare AI has emerged as one of the strongest areas of vertical-AI investment and adoption, and independent research on what works in clinical practice increasingly aligns with the architectural choices Medicus Tiro made early. The market timing and the design thesis converge here.

8 unicorns

Healthcare AI is one of the strongest vertical AI sectors — more than legal, financial services, and media combined.1 Menlo Ventures, 2025

85%

Of healthcare AI spend goes to startups. AI-native architecture is beating bolted-on AI in the category where incumbents had every structural advantage.1 Menlo Ventures, 2025

Narrow wins

Independent research finds adoption favors narrow, task-specific systems — which is why our first product enters through one specialty rather than the whole decade at once.2 ARISE Network, 2026

52→39%

Clinician burnout fell in a large ambient-documentation deployment. Where the documentation burden fell, burnout fell with it — and this burden accumulates at every stage of the decade.4 JAMA Network Open / Mass General Brigham, 2025

The market data in full (Menlo Ventures, 2025)
Market · Category Breakout

Healthcare AI is one of the strongest vertical AI sectors.

Eight unicorns — more than legal, financial services, and media combined. $1.4 billion in 2025 spend, up roughly 3x year-over-year. Among the categories tracked, only horizontal chatbots and coding assistants were reported to be expanding faster. The dollar flow follows the adoption: healthcare is deploying AI at 2.2x the rate of the broader economy.

1 Menlo Ventures, 2025: The State of AI in Healthcare, October 2025.

Market · AI-Native Wins

85% of healthcare AI spend goes to startups.

In ambient scribing — the category where incumbents had the largest head start — AI-native challengers like Abridge and Ambience have captured nearly 70% of the new market, despite Nuance’s 77% pre-AI deployment in U.S. hospitals. Bolted-on AI is losing to AI-native architecture in the category where it had every structural advantage.

1 Menlo Ventures, 2025: The State of AI in Healthcare, October 2025.

What the independent research says about the architecture (ARISE, 2026)
Architecture · Narrow Wins

FDA clearance is increasing, but near-term clinical adoption will favor narrow, task-specific systems. AI tools that are tightly scoped to specific domains and contexts are more likely to demonstrate value and be adopted in practice.

Why this matters for Medicus Tiro. Entering through one specialty isn’t a positioning choice — it is the architecture the research points to. Specialty-specific curated data sets, milestones, and note conventions: tightly scoped, deeply built, with additional specialties following on the same foundation.

2 ARISE Network, State of Clinical AI 2026, Top Takeaway #6.

Architecture · Beyond Scribing

Workflow tools like AI scribes feel transformative, yet objective gains are still modest. The addition of downstream workflow tasks will likely yield more productivity and efficiency impact.

Why this matters for Medicus Tiro. ARISE independently describes the move our capture architecture makes: capture the moment once, then prepare the downstream work it feeds — evaluations, attestations, billing, milestone evidence, competency tracking — rather than the chart note alone. The architecture the research points toward is the one already in build.

2 ARISE Network, State of Clinical AI 2026, Executive Summary.

The burden-to-burnout evidence, procurement priorities, and full source list (1–4)
Architecture · The Burden-to-Burnout Lever

Where the documentation burden falls, burnout has fallen with it.

The clearest real-world signal of what ambient AI delivers isn’t faster typing — it’s relief. Large deployments and controlled studies associate ambient documentation with lower documentation burden, reduced cognitive load, less after-hours work, and measurably lower clinician burnout. That lever exists for a single workflow; the physician-development decade, where the burden has compounded for 25 years and repeats at every transition, is a far larger surface for it.

10,000+
Physicians & staff
Reduced documentation burden across a large-scale ambient deployment.
4 Kaiser Permanente, 2024
2.64 pt
Lower cognitive load
Measured reduction in clinician cognitive load after deployment.
4 JAMA Network Open, 2025
0.90 hrs
Less after-hours work / wk
Fewer after-hours hours per week spent on documentation.
4 JAMA Network Open, 2025
52→39%
Clinician burnout
Burnout fell from 52% to 39% — a 21% relative reduction.
4 JAMA Network Open / Mass General Brigham, 2025

Why this matters for Medicus Tiro. These are reductions in subjective burden, cognitive load, and burnout — the relief clinicians report when documentation work is lifted, not a claim about transcription time saved. That is the lever MomentReady pulls, applied to the administrative work of training rather than a single chart note. Where the documentation burden falls, burnout has fallen with it — and across this decade the burden never stops accumulating.

Procurement priorities in healthcare AI now run maturity, then risk to patient care, then short-term value, with cost secondary.3 Variable Autonomy is the architectural answer to that second priority. We built for the buyer the market is becoming.

Sources · numbering used throughout this page

1 Menlo Ventures, 2025: The State of AI in Healthcare, October 2025. Survey of 700+ healthcare executives. 2 Brodeur PG, Goh E, Tat E, et al., State of Clinical AI 2026, ARISE Network, January 2026. Targeted review of high-impact clinical AI literature.

3 Per Menlo Ventures, leading health systems (Mayo Clinic, Cleveland Clinic, Kaiser Permanente) prioritize, in order: (1) maturity of technology, (2) level of risk to patient care, (3) short-term value delivery. Cost is described as secondary.

4 Ambient documentation outcomes: Kaiser Permanente, 2024 (large-scale deployment, reduced documentation burden); JAMA Network Open, 2025 (cognitive-load and after-hours reductions); JAMA Network Open / Mass General Brigham, 2025 (burnout 52%→39%). Figures reflect reductions in subjective burden, cognitive load, and burnout, not raw transcription time saved.

Next Planned after GME Manager pilot validation

Four more agents. One foundation.

These four agents extend GME Manager inside residency — each planned to take a workflow currently running on tools that weren't built for today's residency programs and rebuild it on the same agent. Build follows commitment: development advances when a committed pilot or design partner is in place.

Partner-dependent · Build follows commitment
📋

GME Onboarding Manager

A planned onboarding AI configured from your specialty, state, and class list, supporting residents from Match Day through Day One—including hStream ID provisioning and state license intelligence.

Partner-dependent · Build follows commitment
✅

GME Recruiting Manager

Composable agentic AI for residency recruitment. One-Click Applicant Profiles™ turn ERAS data into reviewable summaries across screening, interviewing, and ranking, with Variable Autonomy™ at every stage.

Partner-dependent · Build follows commitment
⏰

GME Schedule Manager

A shift scheduler purpose-built for GME, designed to enforce 8 ACGME duty-hour rules in real time with five AI engines on each action: compliance, wellness, training fit, fairness, and supervision. Block tools schedule blocks; hospital tools schedule shifts—a gap general-purpose platforms do not serve well.

Partner-dependent · Build follows commitment
💰

GME Reimbursement Manager

A planned agent whose value is intended to show up in dollars, not just hours saved. Designed to help capture the CMS codes introduced in 2024 and 2025—G2211, APCM, and GC/GE teaching physician modifiers—inside the same system that tracks supervision and procedures.

Design partners shape what gets built next.

Across the GME agents above and the decade-wide portfolio further up this page, qualified organizations can discuss committed pilots or design partnerships that provide the validation, operational context, and commitment required before development advances.

Discuss a Design Partnership
Founder-Market Fit

Not a pivot. A career capstone.

Thirty years of healthcare operating experience—now aimed at the physician journey.

Michael Sousa, Founder of Medicus Tiro and GME Manager

Medicus Tiro was founded by Michael Sousa, a healthcare technology executive with 30 years applying technology to workforce, credentialing, privileging, onboarding, compliance, scheduling, and professional development.

He spent 10 years at IBM and 20 at HealthStream, where he served as an Executive Vice President and President, Credentialing & Scheduling. In 2026 he completed Stanford Medicine's AI in Healthcare Leadership & Strategy program.

Burnout is the symptom. Twenty-five years of accumulated administrative burden is the cause.

Why this became the company

That experience revealed a recurring pattern across the physician journey: administrative requirements accumulate, but the infrastructure needed to perform the work does not.

IBM — 10 years
Higher education and U.S. healthcare leadership
HealthStream — 20 years
Executive Vice President and President, Credentialing & Scheduling
Stanford Medicine — 2026
AI in Healthcare Leadership & Strategy
Founder & CEO, Medicus Tiro and GME Manager
AI-native infrastructure for the physician journey

Thirty years inside the infrastructure that supports healthcare professionals—now rebuilding the least supported decade of the physician journey.

Investor Relations

Building the operating system for physician careers

A $1.57B total physician-journey market—reached from a $183M core GME platform, one credentialing cycle at a time.

$1.57B
Total Physician-Journey Market
$183M
Core GME Platform
35+
Year Customer Relationships
How the $1.57B unfolds
$183M
Entry Point (Core GME)
GME Manager Core Platform + Agent Studio — our first product. ACGME residents and fellows at $840/year, plus non-ACGME, international, data licensing, and premium analytics.
$1.39B
From 15 More Segments
Clinical moment capture & AI workflow ($253M), transition-to-practice & alumni network ($167M), APP oversight AI ($150M), precision learning & test prep ($125M), applicant platform ($115M), and 10 more.
$1.57B
Total Physician-Journey Market
16 segments across the least supported decade, from the first year of medical school to the first year of independent practice.
Strategic Investors
HealthStream — Investor & Integration Partner

HealthStream made a $500K strategic investment and is the sole strategic investor in the GME Manager pre-seed round: a NASDAQ-listed workforce-software leader backing a former EVP. Not a financial bet, a strategic one.

$500K
Strategic investment
Partner
Strategic partnership
Live
hStream integration

Interested in learning more?

Medicus Tiro is building AI that performs the administrative work of the physician-development decade, reaching the broader portfolio through a shared clinical-moment architecture, specialty-specific data, and a disciplined land-and-expand strategy.

What investors are evaluating

A built operating foundation, a six-program Family Medicine pilot initiated in July 2026, a $500K strategic investment and partnership with HealthStream with a live hStream integration, a run-alongside go-to-market model, a shared architecture, a broader product portfolio, and founder-market fit.

Contact Investor Relations

Market and company information as of August 2026

Start a Conversation

Find the right way in

For investors evaluating the full decade, organizations ready to commit to what gets built next, and residency programs adopting our first product now.

Now · Built · Pilot Initiated

Residency Programs

See how GME Manager runs alongside MedHub, New Innovations, or your current residency-management system, starting with procedure logging and AI-drafted evaluations. No migration required, a No-PHI design, and your incumbent platform can remain the system of record. The Phase 1 cohort is fully allocated; the next-opening waitlist is open through GME Manager.

Join the Next Pilot Cohort
Next & Future · Partner-Dependent

Design Partners

Build follows commitment. Qualified organizations can discuss committed pilots or design partnerships for the planned GME agents and the broader Medicus Tiro opportunities, providing the validation, operational context, and commitment required before development advances.

Discuss a Design Partnership
Investor Relations

Investors & Strategic Partners

Discuss the company, current traction, and the expansion strategy: a built operating foundation, a six-program Family Medicine pilot initiated July 2026, a $500K strategic investment and partnership with HealthStream, and a $1.57B physician-journey market entered through a $183M core GME platform.

Contact Investor Relations
Get in Touch

Let's talk

Medical schools, residency programs, APP training programs, investors, and potential design partners—we'd love to hear from you.

✉️
General Inquiries
michael@medicustiro.ai
📅
Request a GME Manager Conversation
Request on GMEManager.ai