Healthcare & health tech

Healthcare AI engineering, led by a working health-tech CTO.

Foundry 41 builds EHR integrations, patient-facing apps, and human-approved AI workflows for health-tech founders, digital-health startups, and practices, clinics, and MSOs. The work is led by Anthony, CTO at a healthcare technology company, who also serves as a fractional CTO for healthcare and AI SaaS companies.

PHI stays in your environmentBAA before any PHI accessA person approves consequential steps

Last updated

Who this is for

Which healthcare teams does Foundry 41 work with?

Three kinds: founders building a health-tech product who need senior technical leadership, digital-health startups that need EHR integrations and AI features shipped carefully, and practices, clinics, and MSOs that want administrative work prepared by AI and approved by staff.

HEALTH-TECH FOUNDERS

A fractional CTO who still writes code

Architecture, build-versus-buy and EHR integration strategy, security and vendor-review readiness, hiring plans, and hands-on engineering when that is the fastest way forward.

DIGITAL-HEALTH STARTUPS

Integrations and AI features that hold up

EHR integration adapters, FHIR APIs, patient-facing web apps on AWS, and AI features with evaluation and human review designed in from the start.

PRACTICES, CLINICS & MSOS

One administrative workflow at a time

Patient intake, prior authorization, or chart audits—prepared by AI, approved by your staff, and running in your environment.

Outside healthcare, the same team does product design and development, AI engineering and company brains.

Anthony, founder of Foundry 41

Who you'll work with

Who is Anthony?

Anthony is CTO at a healthcare technology company and the founder of Foundry 41. He works as a fractional CTO for healthcare and AI SaaS companies and is Expert-Vetted on Upwork, with more than $200K earned there.

His work centers on the parts that decide whether healthcare software ships: EHR integrations, data flows that respect PHI boundaries, AWS environments with narrow access, and AI features a clinician, coder, or front-desk lead can check before anything changes.

Foundry 41 is a small, senior engineering boutique led by Anthony. The work is senior-led and AI-assisted, and you work directly with the people building it.

About Anthony

Current role
CTO at a healthcare technology company
Fractional CTO
Healthcare and AI SaaS companies: architecture, roadmap, vendor and integration decisions, and engineering leadership.
FHIR
Reading and writing FHIR resources through EHR and platform APIs.
EHR vendor APIs
Vendor APIs, including the partner or developer programs each EHR requires.
AWS
Account structure, IAM least privilege, encryption, logging, and environments that keep PHI inside the client's boundary.
AI workflows
Agents that prepare work, evaluations that check it, and approval gates before anything consequential happens.

Use cases

What can AI do in a healthcare workflow?

AI can prepare the repetitive, rule-bound parts of healthcare work—collecting, checking, drafting, and flagging—so staff spend their time on decisions. In every use case below, the agent prepares and a person approves.

Illustrative scopes. Each is defined in a Workflow Blueprint against your systems, volumes, and rules.
Use caseWhat the agent preparesWhat a person decidesSystems usually involved
Patient intakeChecks intake forms for missing or inconsistent fields, matches insurance details, and drafts follow-up requests to the patient.Front-desk staff confirm and send. Nothing writes to the chart without approval.EHR or practice-management system, intake forms, eligibility checks
Prior authorizationAssembles the packet from the chart and the payer's rules, flags missing documentation, and drafts the submission.Staff or the clinician reviews, corrects, and submits.EHR, payer portals or APIs, document storage
Documentation and coding auditsCompares each day's charts with the CPT codes that were billed and lists likely mismatches with the supporting note text.A coder or biller works each flag and decides on any correction.EHR notes, billing and claims data
Care navigationDrafts next-step outreach, reminders, and referral follow-ups from the care plan and schedule.A care coordinator edits and approves each message.EHR, patient messaging, scheduling
EHR integration adaptersEngineering work: adapters that move data between your product and EHRs over FHIR or vendor APIs, with retries, logging, and reconciliation.Your team approves field mappings and each release.FHIR endpoints, EHR vendor APIs
Patient-facing web appsEngineering work: intake, scheduling, education, and portal-style apps on AWS, connected to the EHR where needed.Your team approves content, access rules, and releases.Your product, EHR integration, AWS

An example workflow

What does a daily chart audit look like?

Each morning the agent compares yesterday's documented visits with the CPT codes that were billed and prepares a review list. A coder works the list; nothing in the EHR or the claim changes until they approve.

CHART AUDIT · ILLUSTRATIVE · FICTIONAL DATA

One review list, one owner, every morning.

The agent reads, compares, and explains. The coder decides. The record keeps the note excerpt, the billed code, the flag, and the decision together.

46ENCOUNTERS READ
5FLAGGED FOR REVIEW
0CHANGES WITHOUT APPROVAL
1CODER
AUDIT RUN · ILLUSTRATIVEOWNER · CODING LEAD
01
Read yesterday's encounters

Read-only access to signed notes and the codes submitted for billing.

COMPLETE
02
Compare documentation with billed codes

Each flag cites the note text and the rule it was checked against.

COMPLETE
03
Coder reviews the flags

Accept, dismiss with a reason, or send back to the clinician.

REVIEW REQUIRED
04
Approved corrections go to billing

Only after approval; dismissals feed the next rule review.

PENDING

PHI, BAAs, and control

How is PHI handled?

PHI stays in your environment. We sign a Business Associate Agreement (BAA) before any PHI access, work through delegated access you can revoke, and design each workflow so the agent sees only the minimum data the task needs.

We build compliance-minded architecture; we do not certify HIPAA compliance. Your compliance program, policies, and vendor reviews determine that, and we support their review with data-flow diagrams, access matrices, and run records.

Talk through the architecture
  • 01
    BAA before PHI

    No PHI access until a BAA is signed. Before that, we work with synthetic or de-identified data.

  • 02
    Your environment, your keys

    Systems run in your cloud account. Logs, records, and outputs stay with you.

  • 03
    Minimum necessary access

    Scoped, revocable credentials per workflow, read-only unless a write step is approved in the Blueprint.

  • 04
    Approved providers only

    PHI goes only to model and API providers you have approved under your own agreements.

  • 05
    A person approves

    Anything that writes to the chart, changes a claim, or messages a patient waits for a named person.

  • 06
    A record of every run

    Inputs, outputs, flags, decisions, and approvers are kept with the work.

Working together

How does a healthcare engagement start?

With a short fit call. From there it is either a fractional CTO engagement, or a paid, fixed-scope Workflow Blueprint for one workflow, followed by a build and validation targeted at 30 business days from Blueprint kickoff.

The 30-business-day window is a design-partner target that depends on access, integration readiness, and your security review; it is not a promise. We do not publish prices; scope and fee are agreed before any paid work starts.

Questions

Frequently asked questions about healthcare work

Do you offer fractional CTO services for healthcare startups?

Yes. Anthony, CTO at a healthcare technology company, works as a fractional CTO for healthcare and AI SaaS companies: architecture, roadmap, EHR integration and vendor decisions, security and vendor-review readiness, hiring, and hands-on engineering.

Which EHRs and standards do you integrate with?

We build against FHIR APIs and EHR vendor APIs. Access programs differ by vendor, so the Workflow Blueprint confirms what each EHR allows before a build.

Will you sign a BAA?

Yes. We sign a Business Associate Agreement before any PHI access. Until it is signed, we work only with synthetic or de-identified data.

Is your work HIPAA compliant?

We do not claim or certify HIPAA compliance. We build compliance-minded architecture: PHI kept in your environment, least-privilege access, audit records, and human approval. Your compliance program, policies, and vendor reviews determine compliance, and we support that review.

Where does PHI live?

In your environment. Systems run in your cloud account under your keys, PHI goes only to model and API providers you have approved under your own agreements, and logs and run records stay with you.

Can the AI change the EHR or a claim on its own?

Not in the workflows we build by default. The agent prepares drafts, flags, and packets; anything that writes to the chart, changes a claim, or messages a patient waits for a named person's approval. Autonomy levels are set in the Workflow Blueprint and change only with the workflow owner's approval.

Which healthcare workflows make good first projects?

Repetitive, rule-bound work that crosses systems and already has a person who checks it: patient intake, prior authorization packets, daily chart audits against billed CPT codes, care-navigation follow-ups, and EHR integration adapters.

Have you built this for healthcare customers before?

Anthony builds healthcare software as CTO at a healthcare technology company. Foundry 41's own AI workflow references, Sky and Olivia, are internal deployments for our own companies, not healthcare customer deployments, and we do not show client work without the client's permission.

How do we start?

Email hello@foundry41.com for a short fit call. Tell us the workflow or technical decision, the systems involved, and who owns the outcome. The call decides whether a fractional CTO engagement, a Workflow Blueprint, or neither is the right next step.

Bring one workflow or one technical decision

Start with a fit call.

Tell us the workflow, the systems involved, and who owns the outcome. A short call tells both sides whether it is a fractional CTO engagement, a Workflow Blueprint, or not a fit.