HayanMind presents FreeDoc

Exploring local-first AI for medical claim administration.

FreeDoc is an early-stage concept for supporting claim administration alongside the EMR clinics already use. We are exploring how local AI could interpret visible screen context, surface relevant guidance, and prepare administrative next steps for human review.

Discuss an Evaluation Explore the Concept
  • Existing EMR context
  • Local-first design
  • Human review
Illustrative FreeDoc interface concept alongside an existing EMR
Human review point Screen context Rule guidance

Conceptual workflow No patient data is shown.

The administrative gap

Claim administration can involve fragmented screens and manual rule checks.

The work can be repetitive and highly specific to each site. Depending on the local workflow, staff may move between legacy interfaces and detailed claim guidance, then review the same information again before submission.

  1. 01Different interfaces

    Legacy EMR screens and workflows vary between providers and sites.

  2. 02Detailed criteria

    Insurance claim rules can be detailed and may change over time.

  3. 03Repeated handling

    Lookup, entry, review, and correction may add to administrative workload.

Potential administrative burden

Depending on the site, claim review may involve repeated lookup, entry, checking, and correction across different systems.

Proposed workflow

Same screen.
Assisted workflow.

The concept is intended to support selected steps within the visible workflow, with a person retaining responsibility for review and approval.

  1. 01Observe

    Use the visible EMR screen as context.

  2. 02Interpret

    Identify potentially relevant fields and workflow state.

  3. 03Reference

    Surface potentially applicable guidance for review.

  4. 04Prepare

    Prepare a possible next administrative action.

  5. 05Review

    Require human review before any operational use.

Boundary FreeDoc is an early-stage administrative support concept. It is not presented as an autonomous system for clinical judgment, diagnosis, treatment, or claim submission.

Why FreeDoc

Designed to explore a path alongside existing systems.

01

Screen-level approach

The initial concept explores working through the visible interface rather than depending on early EMR API integration.

Technical feasibility, configuration needs, and workflow fit would still require validation at each site.
02

Local-first design

The proposed architecture is intended to keep core model inference on a dedicated clinic-side device.

This design does not by itself establish privacy, security, or regulatory compliance. Controls would need partner-specific review and validation.
03

Human review by design

The concept is intended to present possible administrative next steps while keeping clinicians and administrators responsible for decisions.

Any operational use would require defined approval points, exception handling, and local validation. No regulatory certification is claimed.

Core architecture

Keeping the model close to the workflow.

The proposed architecture would place core model inference on a dedicated clinic-side device, alongside the EMR already in use. This is a design goal that would need validation in each partner environment.

Concept visualization
Local intelligence. Human control. Conceptual workflow only. No patient data is shown.
  1. InputExisting EMRUse the visible interface as context
  2. InferenceLocal LLMIntended to run the core model near the workflow
  3. SupportRule guidanceSurface guidance for review
  4. ControlHuman reviewKeep an accountable operator in the loop

Access, logging, retention, masking, security controls, and jurisdiction-specific requirements would need to be defined and validated with each partner.

Proposed safeguards

Checks are part of the design.

The current concept explores how screen context, rule guidance, verification, and human review could be combined in a controlled administrative workflow.

  1. 01Screen contextObserve the visible workflow
  2. 02Task contextInfer a possible administrative goal
  3. 03Rule referenceSurface potentially relevant guidance
  4. 04Action previewPrepare a candidate next step
  5. 05Human reviewRequire operator confirmation

Design principle

The concept assumes review before action.

Context checks and an action preview are included as design goals, not as evidence of production reliability.

ObserveUnderstand the visible workflow
ReviewPresent a candidate action for human review

Conceptual architecture. No claim is made about clinical performance, claim accuracy, regulatory compliance, or production readiness. Operational performance would need a limited, workflow-specific evaluation.

Screen contextUse visible EMR elements as input to the prototype.
Rule referenceExplore how administrative guidance could be surfaced for review.
Local-firstTarget core inference on a dedicated clinic-side device.
Human reviewKeep clinicians and administrators responsible for decisions.

The public website describes design concepts only. Workflow details and validation scope are defined with each partner.

Current stage

An early-stage, controlled prototype.

The prototype visuals illustrate how screen context, claim guidance, and human review might work together. No production deployment or jurisdiction-specific readiness is claimed.

Illustrated
Screen context, rule guidance, and human review in a controlled concept.
Data
Synthetic or non-production data only.
Not yet validated
Operational performance, production security controls, or regulatory fit.
Next step
Define a limited evaluation and its boundaries with each partner.
Conceptual FreeDoc review interface showing information, guidance, and a human review point
Prototype concept
Prototype visualization. The interface illustrates a proposed approach, not a live clinical or production system.

Partnership

Let us evaluate one administrative workflow together.

We are looking for partners willing to help define the operating context and assess whether the concept is relevant before any implementation decision.

  • Clinic, hospital group, or MSO
  • EMR, HIMS, or RCM vendor
  • Healthcare IT distributor or implementation partner
  1. 01

    Select a candidate workflowStart with one administrative task worth exploring.

  2. 02

    Define assumptions and boundariesMap risks, exceptions, and human review points.

  3. 03

    Review local requirementsConsider regulatory, privacy, security, and data-handling needs early.

  4. 04

    Plan a limited evaluationUse non-production or shadow-mode conditions where appropriate, without operational submission.

  5. 05

    Learn togetherAssess relevance, exceptions, user feedback, and control requirements.

For Singapore, we are seeking local validation—not presenting a ready-to-deploy product. Any work would begin with local workflow, regulatory, privacy, and security review.

HayanMind

A KAIST-rooted software and AI team based in Daejeon.

HayanMind is applying its software product, partner collaboration, and AI experience to this early-stage concept.

Product craftExperience designing and operating software products.
Applied AIA practical approach to AI-enabled workflows.
Partner-ledCollaboration around local workflow and operating requirements.

HayanMind's broader software experience should not be read as evidence of FreeDoc deployment, performance, or regulatory readiness.

A narrow first evaluation

Let us evaluate one administrative workflow together.

A limited evaluation could help determine whether the concept fits the local workflow and what data, security, approval, and regulatory boundaries would be needed.

Discuss an Evaluation
Jay Oh CEO, HayanMind Inc. jmoh@hayanmind.com

Do not include patient or medical information in an email to us.