Technology and trust
Trust should be designed in, not added later.
Review ASKLEMER’s proprietary SOAP data, consent, de-identification, controlled training, local-inference roadmap, and practitioner safeguards.
ASKLEMER is being designed as a controlled clinical-data loop: authorized proprietary SOAP-note data supports model development, practitioners review every output, and future local inference is intended to minimize routine cloud transmission.
Proprietary SOAP corpus
The clinical foundation is a proprietary corpus of real-world acupuncture SOAP notes from educational and practice-clinic settings—not synthetic dialogue alone.
Consent before reuse
Future patient records enter the improvement pipeline only after affirmative opt-in under an appropriate authorization process.
De-identify first
Authorized information is de-identified before it is transferred into the proprietary training environment.
Controlled training
Model development is intended to use infrastructure ASKLEMER controls or deliberately selects; customer data is not for training third-party foundation models.
Local inference target
Updated models are intended to run on practitioner-selected local inference units so routine clinical inference can remain on premises.
Practitioner in the loop
The system assists, drafts, and suggests. It does not finalize a note or determine a clinical plan without practitioner review.
Closed-loop lifecycle
From opt-in data to local model updates.
Authorize
The patient affirmatively chooses whether eligible future chart information may support model improvement.
De-identify
Authorized clinical information is separated from identifying information before training use.
Train privately
De-identified data enhances models within owned or deliberately selected infrastructure, including controlled data-center capacity.
Propagate locally
Validated model updates are intended to return to local inference units selected by practitioners.
Infrastructure boundary
Selected infrastructure is not an open data pipeline.
ASKLEMER may operate training workloads in its own facilities or rent capacity from selected infrastructure providers such as AWS. That infrastructure is intended to operate under ASKLEMER’s controlled environment, contracts, access policies, encryption, and audit controls.
The local-inference roadmap is a target architecture. Until it is fully implemented and validated, the site will not claim that every workflow is cloud-free or that completed certifications exist.
Access paths
Standalone or integrated with the practitioner’s EHR.
Standalone website: a dedicated clinical workspace for structured SOAP drafting, cited retrieval, and practitioner review.
EHR integration: planned connections to compatible systems such as athenahealth through supported APIs, command-line tooling, or MCP interfaces. Access, write-back, and partner permissions must be confirmed before use.
Product boundaries
- ASKLEMER is in development and is not a medical device.
- It does not provide medical advice to patients.
- It does not diagnose, treat, or cure.
- Its outputs require licensed-practitioner review.
- Security commitments are design principles, not completed certifications.
- Privacy and compliance obligations depend on final deployment and use.
Have a diligence question?
Ask about the intended controls, evaluation plan, or pilot safeguards.