01
Decision Intelligence Systems
Turning data, models, and signals into structured, explainable decisions — not raw predictions left for someone else to interpret.
We design AI systems that show their reasoning.
Most AI in production hands you an answer with no account of how it got there. We build the layer that makes the assumptions, the uncertainty, and the logic visible — so people can accept, challenge, or override what the system concludes.
Get in touchWho we are
Lanterns.fun is an independent practice working on decision intelligence and AI reasoning. We work with organizations putting AI into workflows where the outcome matters and someone has to answer for it.
Our starting position is that automation should not cost you your ability to explain a decision. As AI moves further into operational work, reasoning tends to disappear into opaque models and fragmented processes. We design against that — keeping human judgment central, at scale.
What we do
Four capabilities, usually delivered together.
02
AI Reasoning Architecture
Multi-agent reasoning frameworks that define how a system reads its inputs, resolves conflicting signals, and reaches a conclusion you can trace back.
03
Cognitive Workflow Design
Human–AI workflows with real oversight: where a person reviews, what triggers escalation, and how domain expertise enters the loop.
04
Judgment Risk Modeling
Mapping failure modes, bias, and how uncertainty compounds through a decision — before deployment, not after an incident.
Working on something where the reasoning has to hold up?
Tell us what you're deploying and where you need it to be defensible.
Start a conversation