Standards authority becomes market power when it becomes executable.
AI is starting to help with decisions about people's health, in hospitals, and in pro sports, where teams use data about players' bodies to decide who plays and when. Darwin by Medigram is the system that checks the AI: it scores every decision on six kinds of trustworthiness and keeps a permanent record, so whether the stakes are a patient's care or a player's career, nothing happens unchecked and there's always proof. Designed to work on mobile, where the work happens.
Here is the plain version: most companies that help write AI safety standards do not also build the product and prove it behaves as promised. Medigram does all three: write the standard, build to it, and independently verify it. That closed loop is what is starting to show up in how hospitals and buyers make purchasing decisions.
A short reader path into Building Darwin, the full CEO Letter.
The investment thesis
AI governance is moving from policy to infrastructure. IEEE/UL 2933:2024 with TIPPSS (trust, identity, privacy, protection, safety and security) has reached contracting and CIO guidance. Darwin implements the governance discipline its architect helped write, then tests whether the resulting controls actually work.
Why this is difficult to copy
The scarce capability is the loop. Standards interpretation, architecture, AI engineering, security boundaries, evidence design and assurance normally live in different teams or institutions. Medigram has demonstrated those layers in one operating system and maintains a claims register tying assertions to verification.
Why the loop closes here
In plain terms: writing a safety standard, building the product that follows it, and proving the product actually behaves are normally three different jobs, done by three different institutions, and something gets lost every time the work changes hands. That matters to an investor because a loop that does not lose information end to end is hard to copy and hard to fake. Medigram’s version of that loop (standard, to implementation, to verification, to evidence, to standard evolution, as this letter describes it) closes here not because of an org chart, but because of one person’s range.
Why Sherri and the Medigram team
Sherri Douville is not only the person who helped write the underlying standard. She runs Medigram day to day, personally does six distinct technical jobs inside it (standards authorship, architecture, AI engineering, security engineering, data architecture, and assurance engineering), and on top of both of those does the work of keeping the wider trust ecosystem functional: she co-founded and chairs the Trustworthy Technology & Innovation Consortium (TTIC), built to convene the leaders across medicine, engineering, security, and law, because, as TTIC’s own account of its founding puts it, “each perspective is necessary. None is sufficient on its own.”
That is three jobs stacked, not one: running the company, doing technical work most organizations split across six people, and sustaining the standards ecosystem the whole approach depends on. She personally holds the standards seats behind that ecosystem (IEEE/UL 2933 Trust and Identity Subgroup Co-Chair, ANSI/HSI 2800-2025 working group member) and is sole committer of record across 223 commits on Darwin’s implementation, directing agentic coding rather than hand-coding. That is evidence that, today, those three jobs are one chain of judgment rather than three separate claims. The investment question is whether that discipline becomes Medigram’s institutional capability, repeatable without her personally carrying all three, not whether the company can keep depending on one person to hold them together indefinitely. TTIC's own relationship disclosure and continuity page addresses this directly: the practice is encoded in standards, a published definitional corpus, and the TTIC and AIMed faculty networks, not held by one person alone.
What compounds
Standards inform implementation. Implementation reveals whether requirements are technically workable. Behavioral testing finds failures. Remediation and regression testing produce evidence. Evidence improves the system and informs future standards work.
What to diligence
The central diligence question is whether this unusual concentration of standards, architecture, implementation, verification and evidence capability can convert into repeatable institutional adoption. Qualified counterparties can review supporting chronology, claims evidence and withheld technical materials under NDA.
Need the full architecture and evidence?
The CEO Letter contains the detailed technical rationale, verification loop, standards context and diligence path.