AI lab & experiments

AI applied to healthcare problems I have actually seen.

These are labelled honestly. Nothing here is presented as a shipped product unless it is — status badges tell you exactly where each idea stands.

Dental AI Assistant

PROTOTYPE

The problem

Dental practices answer the same patient questions hundreds of times a week, and clinical staff lose hours to triage that is not clinical work.

The idea

An assistant grounded strictly in a clinic's own approved information that handles pre-visit questions, explains treatments in plain language, and escalates anything clinical to a human.

Intended capabilities

Grounded answers from clinic-approved content only
Plain-language treatment explanations
Explicit hand-off for clinical questions
Refuses to give diagnosis or medical advice

Stack direction

LLM APIRetrieval over curated contentWeb UI[EDITABLE STACK]

Healthcare GRC AI Assistant

CONCEPT

The problem

Governance, risk and compliance work in healthcare is document-heavy and repetitive, yet mistakes carry real consequences.

The idea

An assistant that helps teams navigate their own policies and evidence — surfacing the relevant clause with a citation instead of a confident guess.

Intended capabilities

Answers cited back to the source document
Policy and control lookup
Draft support for recurring documentation
Abstains when the source material does not cover the question

Stack direction

LLM APIDocument ingestionCitations layer[EDITABLE STACK]

AI Marketing Assistant

CONCEPT

The problem

Healthcare marketing has to stay accurate and compliant, which is exactly where generic AI content tools fail.

The idea

A constrained content assistant for clinics that drafts on-brand material inside clinical and regulatory guardrails, with a human approval step by design.

Intended capabilities

Brand and tone constraints
No unverified clinical claims
Channel-specific drafting
Human review built into the workflow

Stack direction

LLM APIPrompt guardrailsReview workflow[EDITABLE STACK]