We build assets that compound
Every brief improves the discipline of source review. Better foundation models help the input layer; the durable work is the review standard and the evidence record around it.
Talent
Build decision timing for medical knowledge change.
NextConsensus builds AI systems that track how medical claims change over time and turn that movement into sourced briefs for high-stakes healthcare decisions. The work runs on temporal replay, claim identity tracking, language-change classifiers, provenance systems, and calibration that compounds with use.
The real problem is judgment under moving uncertainty.
Medical knowledge does not update all at once. A trial shifts the evidence base. Specialists begin changing practice. Payers respond unevenly. Guidelines lag. Labels lag. Public reference systems absorb change through argument, revision, and stabilization.
Institutions sit in the middle, trying to make decisions while every surface is moving on a different timeline. NextConsensus exists to make emerging knowledge change legible early enough for experts and institutions to review it before decisions harden.
Too early is reckless. Too late is expensive. The hard part is knowing when uncertainty has changed enough to require action.
What we believe about the work.
We build decision-timing infrastructure
Medical knowledge changes before institutions can safely admit it has changed. The hard problem is recognizing that moment without pretending uncertainty has disappeared.
We build for the decision
Begin with the disputed claim and decision context. The unit of work is a sourced brief, not a stream of undifferentiated updates.
We separate signal from authorization
Evidence movement is not the same as review obligation. The product governs the space between detection and action, without collapsing it.
We measure before we claim
Hindsight does not leak into the assessment. Each read is pinned to a review date — what was knowable then, not what we know now. If the evidence is thin, the brief says so.
Depth over coverage. Precision over generic summaries.
What we choose to optimize — and what we leave to others.
- We choose depth over coverage. Precision over volume. Measurement integrity over growth metrics.
- We are not building "chat with PubMed." The work is judgment under moving uncertainty.
We release infrastructure, not just briefs.
Sourced through Refract, our open-source developer SDK. It turns raw public revision histories into structured timelines. Anyone can inspect it, run it, or build on it.
Buyers get the healthcare decision layer: sourced briefs with caveats, source references, and review context. Developers and researchers get the observation layer that demonstrates how public knowledge change can be structured from revision histories.
This is not for everyone.
You might be a fit if:
- You care more about measurement integrity than shipping velocity.
- You are comfortable saying "the evidence does not support a conclusion here."
- You think the most interesting infrastructure problem in healthcare is the gap between what is changing and when institutions can act on it.
- You can write clearly about uncertainty without hedging into meaninglessness.
- You want to build a new category, not optimize an existing one.
We show review teams when old assumptions can no longer pass unreviewed.
If building decision infrastructure for that gap is more interesting to you than optimizing existing tooling, get in touch.