About

Built by the person who needed it.

Quadfence is led by a practising dentist who made the decision this product exists to inform — and had almost nothing useful to make it with.

Founder

Dr. Soujanya Maddipati, DDS, MPH

Founder & Chief Executive

In December 2022 she bought a dental practice in North Houston. The seller's system showed patient counts, procedure mix and collections, and a CPA went through the books. Those numbers she could verify.

What nobody could tell her was anything about the market. The broker handed over generic census demographics — education, gender split, race. Not how many dentists per capita already competed for those patients. Not the renter and apartment ratio, which determines how fast a patient base turns over regardless of how well you run the practice. Not which clinical services the surrounding practices already offered, and which gaps were open.

She bought it anyway. She has run Mi Smile Family Dental ever since — Medicaid, CHIP, Medicare Advantage and PPO — and sees every patient herself. Then she spent three years buying marketing she could not measure, from agencies paid whether the practice grew or not.

She is the customer. That is the whole reason this exists.

Background Founder
DDSUniv. of Colorado Denver
MPHUniv. of Oklahoma
LicensureTexas
PracticeHouston, since 2022
Payers servedMedicaid · CHIP · MA · PPO
Patients seenpersonally, every one

Public health before dentistry. Population data analysis is the discipline this product automates — and the payer-mix complexity the engine has to model is the one she works inside every day.

The team we are building

Clinical

A physician director

Clinical and operational depth, and the bridge to the adjacent medical-practice market. Two clinicians who are the customer, in two neighbouring verticals.

Commercial

A sales director

Owns onboarding and the agent motion across territories. Texas first, then parallel states — acquisition scales by adding agents, not overhead.

First hire

An engineering lead

In-house ownership of the platform. The platform is built for us under contract by Airadics, Inc., an engineering firm whose principal is the founder's husband. Quadfence owns the intellectual property outright; the engagement has a fixed completion date and a full handover of source, documentation and infrastructure. The relationship is contracted at arm's length and the agreements are available for review.

What Quadfence is

Governed AI for decisions that can be checked.

The engine underneath AiRadics is not specific to dentistry. Multiple AI models from competing vendors reason over the same question and have to agree before anything acts. Rules are compiled into code the engine is checked against. Every decision is logged so it can be reproduced.

Dental is where we prove it — a founder who is the customer, her own practice as the first place we try it, and a market fragmented enough to enter. Medical practices are the near-adjacent follow: the same ontology structure, different labels. Other local services come later, and not before dental and medical are both working.

Concretely: models from competing vendors draft independently and then critique each other, claims are verified against sources and blocked before publication if they fail, and every decision is logged with enough provenance to be replayed.

Each practice runs in its own cloud project. Their raw marketing data lands there and stays there; only aggregated, surrogate-keyed segments cross into ours, which is what lets the network learn without any clinic’s numbers ever reaching another.

In AiRadics that gate sits in front of everything a patient will ever see. The website, the landing pages, the ads, the images and the video are all generated, and none of it goes live until it has cleared FTC advertising rules, ADA guidance, the advertising requirements of the state dental board, and Google’s standard for health content — which is stricter than for most subjects, and is where generated dental copy usually fails.

Twelve provisional patent applications filed and assigned to Quadfence, covering multi-model consensus and deterministic replay, governance compilation, cryptographic provenance and privacy-preserving federation. Prior-art review determines which convert.

Get in touch

If you run a practice, or you want the deck.

We are pre-product and looking for a small number of Texas practices to work with once there is something to work with. Investors: the deck and the operating model are available on request.

hello@quadfence.ai