Infrastructure for legal judgement

Information was the first generation of legal AI.
Agentic workflows are the second.
Decision clarity is the next frontier.

The AI can draft the answer. The judgement call is still yours, and it should stay that way.

Oikea is the decision-clarity layer for legal AI. It turns model output into structured, auditable reasoning workflows. Lawyers think with AI instead of rubber-stamping it, and finish every matter actually knowing their own case.

Model-agnostic. Works alongside Harvey, Legora, and your existing legal stack.

Settlement recommendationContract risk callTermination riskEvidence sufficiency

Example workflows.

How it works

One workflow. Inputs in, a defensible call out that stays yours.

Input
  • Facts
  • Contracts
  • Emails
  • Case law
Output
  • Reasoning tree
  • Missing facts
  • Strongest arguments
  • Confidence score
  • Recommendation
  • Partner sign-off

Every step is shown rather than hidden inside a model. The lawyer can defend the call to a client, a partner, or a court weeks later, and the firm has a sourced record built to meet the professional rules governing AI use.

Reasoning, every step shown
FactsLawArgumentsRisksRecommendationHuman sign-off
Before
  • Fragmented reasoning
  • Inconsistent calls
  • Hidden assumptions
  • Reviewing a black-box answer
After
  • Structured reasoning
  • Auditable decisions
  • Surfaced and weighed
  • Knowing your own case
Why Oikea

The judgement call is the hardest part of legal work.

"Ther rett och skääl icke är, ther kan icke heller wara lagh."

Olaus Petri · Rules for Judges · Sweden-Finland, c. 1540

FI"Mikä ei ole oikeus ja kohtuus, se ei voi olla lakikaan."

EN"What is not just and fair cannot be law."

Sixteenth-century rules for judges already knew where the real difficulty lay. Not in finding the law, but in the judgement call itself. Agentic tools are getting very good at producing the answer. Oikea is built around the part they should never quietly take over: the reasoning, the weighing, the call. So it stays the lawyer's, and stays oikea [right].

Our guiding principles

The standards we hold ourselves to.

Three principles
// Principle 01

The layer above any model

Oikea is model-agnostic by design. It sits above Harvey, Legora, in-house models, and whatever comes next, turning their output into a decision the lawyer understands, owns, and can defend. The infrastructure for legal judgement is a layer, not a model.

// Principle 02

The lawyer's judgement is the point

A wise lawyer is worth more than a good law. Oikea structures the work around the call. The interpretation and the weighing stay with you. The tool sharpens your thinking. It never replaces it.

// Principle 03
85%

Auditable legal reasoning

Sources cited, steps shown, confidence scored, override always available. The only standard worth having for AI in legal practice, and increasingly the one the rules require.

Infrastructure for legal judgement

Oikea sits on top of the tools you already use.

Configured to your practice areas, playbooks, and compliance posture. The same decision-clarity layer, tuned to how your firm actually decides.

Map your workflows
Settlement recommendation
Contract risk call
Termination risk call
Evidence sufficiency review
Conflict-of-interest determination
Privilege & can-we-act check
AI-output sign-off
AI-use compliance review
Client options & recommendation

Example workflows. Every firm's layer is different.

Talk to us

Map your firm's highest-value judgement workflows.

Tell us about your practice. We'll map the three to five decisions where structured reasoning produces the biggest defensibility and speed gains, alongside the AI you already run.

Infrastructure for legal judgement. Deployed firm by firm.