The hard problem is relational
The question is not whether a model can summarize a paper. It is whether a claim's revision history, read at a dated cutoff, tells a team something the committee it is briefing has not yet absorbed.
Talent
Most AI systems summarize what was published. We are building one that shows how the accepted position moved.
NextConsensus reconstructs how medical claims gain support, shed qualifiers, and spread across public sources, and returns a dated source trail for each one. Every record is registered before the outcome is known and built to be checked against it.
Judgment under moving uncertainty.
A new trial is published. A guideline shifts. A competitor changes the relevant standard of care. A subgroup becomes clinically important.
Yet guideline bodies, regulators, and payers move on different schedules. NextConsensus exists to trace each of those transitions as what it is: a procedural act by a named body, on that body's own calendar, with the evidence that preceded it laid out in date order.
Given the sources visible at this cutoff, how far has the accepted position on this claim already moved, and which named body has yet to act on it?
That problem is not reducible to search, summarization, or workflow tooling. It requires clear propositions, temporal integrity, records registered before outcomes are known, and independent review of what happened.
The work is judgment under moving uncertainty.
We build systems that preserve source context, make uncertainty visible, and leave room for expert review. The exact program data and product roadmap are not published on this page.
What makes this hard.
Literature search and summarization are becoming commodities. The hard problem is reconstructing what was knowable at a date, defining what would settle a claim, and reviewing the record without hindsight.
-
Proposition identity across sources and time
Authorities, actions, populations, deadlines, and resolution sources must remain comparable across historical reconstruction and prospective registration.
-
Temporal integrity
Every record is pinned to an evidence cutoff. Later information cannot enter the registered state. Hindsight does not leak in.
-
Resolution and evaluation
A correct-looking narrative is not enough. Outcomes need independent rules, non-events need structure, and any later claim about accuracy needs a denominator it can be traced to.
What we believe about the work.
We build records that can be checked
The core object is not a document or prompt. It is a defined claim, a frozen evidence cutoff, the sources observed at that cutoff, and a rule for what would settle it.
We earn the long term
The work expands only as public evaluation supports it. Claims about performance stay tied to the record that supports them.
We build for expert judgment
NextConsensus does not replace medical, regulatory, legal, or compliance judgment. A source trail is an input to a decision, not the decision itself.
We separate change from authority
Evidence state, registered record, outcome, and enterprise decision remain separate objects with separate authorities.
Who this work attracts.
People who are skeptical of thin copilots but interested in durable systems for evidence, time, provenance, institutional context, and accountable human judgment:
- You think medical AI should be checked against what happened later, not trusted because its explanations sound plausible.
- You care about proposition design, temporal evidence, and provenance.
- You are comfortable building systems where human adjudication is a designed component, not a gap to be engineered away.
- You can distinguish evidence, judgment, outcome, and decision authority without collapsing them.
- You can write clearly about uncertainty without hedging into meaninglessness.
- You want to build systems that make evidence movement and uncertainty easier to inspect.
Applied AI and ML engineers
Provenance-aware retrieval, scientific document understanding, calibrated classification, temporal reasoning, graph retrieval, uncertainty estimation, and evaluation under expert disagreement.
Knowledge-graph and ontology engineers
Authorities, propositions, populations, interventions, evidence states, target rungs, resolution rules, and outcomes need durable representation.
Clinical informaticists and evidence scientists
Systematic review, HTA, pharmacoepidemiology, medical information, guideline methodology, clinical evidence synthesis, and regulatory science.
Regulated product and design leads
Expert workflow, audit trails, uncertainty communication, decision-state design, trust calibration, and interfaces that make reasoning inspectable.
Medical affairs and regulatory operators
People who understand how evidence progresses through expert recognition, procedural movement, guideline action, regulatory action, and coverage policy.
Infrastructure and systems engineers
Source observation, event pipelines, versioned state, tenant-specific logic, provenance preservation, auditability, permissions, and security.
Depth over coverage. Precision over generic summaries.
What we choose to optimize — and what we leave to others.
- We are not building "chat with PubMed." Search and summarization are basic features, not the product.
- We are not building approval or document-control infrastructure. Customers own governance and action.
- We are not presenting unearned performance as a product promise.
- We choose depth over coverage: clear propositions, temporal integrity, and independent review.
We release infrastructure, not just records.
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.
Enterprise teams get the briefings and private application support. Developers and researchers get the observation layer for structuring public knowledge change from version histories.
Not hiring now — but building the pipeline.
If you want to stay in touch about future roles or technical collaboration, the best way to work with us today is:
- Contribute to Refract — the open core. PRs, issues, and docs are the best way to demonstrate fit.
- Join the talent pool — we'll reach out when a relevant role opens.
- Partner inquiry — if you're an org wanting to embed Refract or co-build the healthcare layer.
Method basis
Research threads behind the work.
The literature defines the problem space, not the finished system. The engineering work is temporal evidence reconstruction, proposition normalization, independent resolution, and prospective evaluation.
- The FAIR Guiding Principles for scientific data management and stewardship
Supports the need for durable, inspectable, reusable evidence and provenance records.
- How citation distortions create unfounded authority: analysis of a citation network
Shows how citation patterns can make a claim appear more settled than the source record supports.
- Guidelines for Human-AI Interaction
Supports interface patterns that keep uncertainty, user control, correction, and human authority visible in AI-assisted systems.
- Living systematic review: 1. Introduction-the why, what, when, and how
Frames evidence synthesis as an updating problem, not a one-time search problem.