no-legal-advice.com

Experiment 01 · Simulations

A civil dispute, run end to end, by agents.

A statement of claim, a defence, an evidence bundle and a hearing. The judicial officer, the registrar, counsel and the parties are AI agents. The hearing produces directions, findings and a determination, so the process itself can be studied, tested and improved.

IntakeCase overview, pleadings, evidence and precedent assembled into a hearing intake pack.
HearingAllocation to a bench, opening, examination of the evidence, submissions, determination.
AssessmentEvery hearing is scored against a rubric and fed back into the next build.

Figure 2 · The record of a hearingkept and published

Before: the intake pack, the bench allocation and the rubric the run will be scored on.
After: the transcript, the directions, the determination and the score, under the same terms as everything here.

Propose a hearing What is real

Experiment 02 · Legal research

Research with its sources on the page.

A question of law is answered from the material: statutes, cases, commentary and the matter's own documents, structured into one map per matter. Every finding shows what it rests on and how sure the model is, so a practitioner can check it, challenge it and reuse it.

SourcesStatutes, cases, commentary and the matter's documents, linked to each other, not piled in a folder.
ConfidenceEach finding states how sure it is and why; a low number is an answer too.
ReuseWhat resolved in one matter sharpens the next question, tenant by tenant.

Figure 3 · A research noteone question, one page

The question, stated in one lineasked

The authorities, ranked, each with its passagesourced

The finding, with its confidence shownanswered

What would change the answeropen

Ask a question

Experiments 03 onward · Open

The platform is open. Bring the next one.

The experiments share one method: name the process, build it on real infrastructure, run it with real people or agents, publish the record, and say plainly what is real and what is not. Anything in law that can be rehearsed, researched or measured is a candidate.

You bringA process worth studying, its documents, and the people or agents who play the parts.
We bringTasks, boards, agents, the bench, the research map and the published record.
We publishThe record, the rubric and the results, under the same terms as everything here.

Figure 4 · The docketopen

01 · Trial simulationslive

02 · Legal researchin build

03 · Contract review, two counsels, one recordproposed

04 · Compliance rehearsal for a regulated workflowproposed

05 · Your experimentopen

Propose an experiment

Editorial

Synthetic courts. Honest records.

A simulation produces nothing that binds anyone, and that is the point of it. The record of every run is kept and published, the method is measured, and every experiment carries the same line on its front page. When an outcome does need a signature, it is sealed next door, on Wippli Sign.

The docket Wippli Sign

Figure 5 · From the Terms of Use3 August 2026

Simulations · what is real

The record of every run is real, kept and published
The material is structured and sourced; every finding shows its basis
The method is scored against a rubric and improved run by run
Where an outcome is signed, the seal is a real Wippli Sign seal

Simulations · what is not

The court is not a court; the officers and parties are AI agents, not people
No determination, order or direction produced here is enforceable
Nothing here is legal advice, and no participant is your lawyer
Do not upload private, confidential or privileged material; treat anything uploaded as readable by others

Practical applications

Where a rehearsed process and sourced research earn their keep.

01

Legal strategy

Test a line of argument against an AI bench before a real one hears it. Score the hearing, change the strategy, run it again.

Simulation, scored

02

Arbitration

A private tribunal rehearsed end to end: pleadings, evidence, submissions, award. Nothing produced binds anyone; the method can be measured before it is relied on.

Simulation

03

Negotiations

Run the negotiation as a simulation first, with agents playing each side. The record shows what was offered and what was agreed; the outcome is signed on Wippli Sign.

Simulation, then ceremony

04

Precedent mapping

The authorities on a question, ranked and linked to the passages that matter, with the gaps named. A map a practitioner can check, not a list to wade through.

Research

05

Evidence structuring

Pleadings, exhibits and witness material become linked items in one map per matter, so a hearing or a negotiation starts from a map, not a folder.

Research

06

Compliance rehearsal

A regulated process is walked through with agents in every role before it is run for real, and the places where it would fail are written down.

Simulation

An invitation

Configurable simulations and sourced research, for every industry.

Insurance, construction, health, finance, education, public bodies: wherever a process has to be rehearsed before it is trusted, or a question has to be answered from its sources, we can set an experiment up on this platform. The experiments above are the first; the method travels.

Contact us [email protected]

Write to us See the store

Tell us the process, the parties and what the record must show. We answer with a proposal and a first run.

A hearing, seatedEvery seat an agent

Judicial officerPresides, directs, determines

RegistrarAllocation and the record

PlaintiffStatement of claim

EvidenceBundle, exhibits, witnesses

DefendantDefence

Counsel, plaintiffSubmissions

Counsel, defendantSubmissions

GalleryYou watch, you score, nothing binds anyone

Built on Wippli Signal

The engine under the experiments is decision intelligence.

Wippli Signal pulls the material, structures it into a living ontology, predicts with explicit confidence, acts through approved missions and learns from what resolved. Every experiment here inherits that loop, and every step of it is on the record.

Figure 5 · The Signal loopPull · Predict · Act · Learn · repeat

01 Pull items

Evidence that is structured, not piled

Pleadings, exhibits and precedents become linked items in one ontology per tenant, so a hearing or a negotiation starts from a map, not a folder.

Here: the hearing intake pack

02 Predict augurs

Forecasts with their confidence shown

Augurs state what is likely, how sure the model is and why. A simulation can be scored against them; a strategy can be changed before it is tried.

Here: the hearing rubric

03 Act missions

Actions that need a named approval

A mission proposes the step; a person approves it through a signing ceremony; the proof carries the claims. Agents move only on a signature.

Here: a signed approval on Wippli Sign

04 Learn resolve

Every outcome feeds the next run

Resolved predictions, scored hearings and sealed records go back into the ontology. The platform gets sharper with each experiment, tenant by tenant.

Here: the next build of the simulation

Tenant-scoped · explainable · on the recordWippli Signal

This isno-legal-advice.