Matt Martin Medware Solutions Pty Ltd · Bondi, Sydney
Submitted to Anthropic Inc.
March 2026 · Confidential
A market opportunity proposal for Anthropic

From liability
to preferred.

The Problem

In regulated industries, "built with AI" is currently a procurement liability — not because the software isn't excellent, but because no trusted verification standard exists. So builders hide it. And adoption stalls.

The Opportunity

Claude Certified is an automated audit framework baked into the build process from day one — turning Claude-built software into the most verifiably trustworthy software in any procurement conversation.

i.

The perception problem
is the entire problem

Procurement departments in hospitals, law firms, banks, and government agencies are not asking "is this software built with AI?" They don't know. They're asking: "Is this software secure? Is it compliant? Can we trust it?"

But the moment they discover it was built by a non-developer using AI tools, the conversation ends. Not because of evidence. Because of narrative. The prevailing assumption is: AI-generated software equals shortcuts, equals risk.

That assumption is wrong — and Anthropic is the only company positioned to prove it at scale.

Today's narrative
"Built with AI"
means cutting corners
With Claude Certified
"Claude Certified Level 3"
means built to a higher standard
Who built this? What level is it certified to?
Prove you're safe Already continuously verified
Non-dev stigma Process superiority
AI = liability to hide Claude Certified = lead with it
Post-build audit, if at all Continuous audit is the build
$10–20K compliance assessment Certification already in hand
ii.

The audit isn't applied
to the build. It is the build.

Traditional development teams audit after they build — if at all, and usually only when forced to. Claude Certified inverts this entirely. The automated auditor runs alongside Claude from the first line of code. Security checks, privacy validation, compliance verification — continuous, logged, traceable. By the time the software ships, the certification is an output of the process itself.

This means something profound for procurement: a solo founder using Claude Certified can produce more verifiably trustworthy software than a twenty-person development team that never ran a continuous audit. The question stops being about who built it. It becomes about what process built it.

And critically: that process only exists inside Claude.

"The certification can only be achieved if you begin with Claude and build within the certified framework."

This is not a badge you apply for after the fact. There is no backdoor. There is no equivalent elsewhere. A competitor cannot certify against this standard by switching models — because the audit trail, the continuous verification, the chain of provenance begins at line one, inside Claude Code, and never leaves.

Claude Certified is simultaneously a trust mark for buyers and an insurmountable moat for Anthropic.

iii.

A tiered certification
architecture

Each level builds on the last. Higher levels require deeper integration with the Claude Certified auditor — increasing the switching cost and the value of the certification simultaneously.

IEntry
Claude Built
Verified Anthropic API usage with basic automated code quality and disclosure standards. Self-assessed. Badge issued on completion.
Verified API usage Code quality baseline AI disclosure standard Self-assessed
Opens doors to SMB and startup procurement
IIPrivacy
Claude Privacy Verified
Continuous automated data handling audit. Verified no training on user data. Mapped to GDPR, Australian Privacy Act, HIPAA. Annual renewal.
Continuous data audit Privacy Act aligned GDPR / HIPAA mapped Annual renewal
Required for any software touching personal data
IIIHealthcare
Claude Healthcare Grade
Clinical workflow validation. Regulatory compliance mapping (PBS, TGA, Medicare, NHS). PSAF-equivalent pre-clearance pathway. Third-party audit required.
Clinical workflow audit Regulatory mapping PSAF pre-clearance Third-party verified
Unlocks hospital and government health procurement
IVEnterprise
Claude Enterprise Certified
Penetration testing integrated into build pipeline. SOC 2 equivalent pathway. Uptime and reliability standards. Continuous monitoring post-deployment.
Pen testing pipeline SOC 2 equivalent Reliability standards Post-deploy monitoring
Unlocks enterprise, finance, and government at scale
iv.

This wall exists right now.

Medflow + Westmead Hospital, Sydney

Medflow automates PBS authority forms for Australian specialist physicians — a genuine clinical efficiency improvement built entirely with Claude, deployed to real clinics, processing real prescriptions.

Westmead Hospital — one of Australia's largest — is interested. The clinical case is clear. The software works. The procurement process cannot move forward without a NSW eHealth PSAF security assessment: $10,000–$20,000 and 6–12 months minimum lead time.

Here is what is not being asked: "Was this built with AI?" Nobody knows. Nobody is asking. They are asking: "Can we trust it?" And right now, the only answer available is an expensive, slow, post-hoc audit that treats every software vendor equally — regardless of how rigorously the software was actually built.

Claude Level 3 Healthcare Grade certification would not replace compliance. It would provide the verified foundation that makes compliance faster, cheaper, and more likely to succeed — and it would exist before the first procurement conversation begins.

Medflow is one company in one country in one vertical. Multiply this by every regulated industry in every market where Claude-built software is hitting the same wall — and the scale of what Claude Certified unlocks becomes visible.

v.

The moment to own
this standard is now

$2.5B
Claude Code annualised
run rate, Feb 2026
29M
Daily VS Code installs
of Claude Code
0
Existing trusted certification
standards for AI-built software

Millions of developers are shipping Claude-built software right now. Governments are writing AI regulation right now. Procurement departments are forming their policies right now. The standard that gets established in the next 12–18 months will be the standard that governs this industry for a decade.

Anthropic can write that standard proactively — or comply with whatever someone else writes reactively. Claude Certified is how Anthropic leads the regulatory conversation rather than responds to it.

vi.

What Anthropic
actually gets

01

Regulated Industry Unlock

Healthcare, legal, finance, government — the largest untapped markets for AI software. Certification is the key. No other approach gets Anthropic into these markets at scale.

02

An Insurmountable Moat

The certification cannot be achieved outside Claude. Competitors cannot replicate it by switching models. Every Claude Certified developer is locked in by the value of their own certification.

03

Regulatory Leadership

The first AI company to establish a voluntary certification framework shapes the standards that regulators will eventually mandate. Anthropic writes the rules, rather than complying with someone else's.

04

Narrative Ownership

Claude Certified flips the AI narrative in procurement from liability to preferred — globally, across every regulated vertical, with Anthropic's brand at the centre of the trust signal.

Matt Martin

Founder & CEO
Medware Solutions Pty Ltd · Bondi, Sydney
Medflow Pty Ltd — PBS authority automation for specialist physicians
Medcast Media Pty Ltd — AI-curated medical audio for healthcare professionals
MedPrep — Patient bowel prep management platform

For consideration by Anthropic's
product and enterprise strategy teams

feedback@anthropic.com
Attn: Mike Krieger, CPO
Dario Amodei, CEO
Anthropic Inc. · San Francisco