← All finance signals
Governance

Machine Mediation

The machine steps between us to settle the dispute, arbitrating contracts, quarrels and the odd relationship.

What is it?

A model takes the seat of first arbiter — triaging claims, drafting settlements, running the early stage of a dispute that used to open with a person. Small-claims platforms and construction arbitrators are the first to hand it real cases. Whether a bound decision follows is still an open handoff.

Why now?

Judgment was always rationed by scarce expert time — judges, arbitrators, adjusters who can only carry so many files. A model reads a contract and drafts a settlement at a caseload no panel matches. The same delegation reshaping ordinary work now reaches the seat where fairness gets decided.

Who’s building it?

Legal-tech vendors and arbitration bodies build the tools; courts and regulators still hold the authority to bind.

ODR platformsArbitration bodiesCourts and regulatorsModel providersInsurers and claims teams
The full read
Why now, who is building it, where the market goes, and what stays unsettled.
Get weekly →
01 · The market
Who’s in it
Builders — legal-tech platforms and the arbitration bodies licensing them.
Buyers — claimants and small businesses first, insurers and corporates once trust holds.
Gatekeepers — courts, bar associations, and the regulators who decide what can bind.
What’s at stake
Where the caseload sits decides who profits — routine claims to software, hard cases still to people.
A bad ruling costs the institution, not the vendor.
Appeal volume tests it — a wave would stall rollout.
What’s unsettled
Whether a party can waive a human decider, or must keep that option.
Who is liable for a wrong ruling — platform, or institution.
Whether speed and access outweigh the loss of a human explaining the reasoning.
02 · How the demand matureda progression, not a timeline
Triage
A model sorts and drafts, cutting the backlog before a human reads it.
Pilot rulings
Small-claims and construction bodies let it decide narrow, low-stakes cases first.
Institutional trust
Accreditation and a clean track record let the caseload widen past the pilot categories.
Binding authority
Courts recognise the ruling as final, not merely advisory, closing the loop.
03 · Where the market goesthe forks, not a prediction
→Advisory or binding — whether courts let a machine ruling stand alone decides if this scales past triage into real judgment.
→Institutions with existing accreditation absorb the tooling fastest — the moat is trust, not the model.
→A contested ruling that draws headlines could freeze adoption well past the case that caused it.
→Value settles wherever the caseload concentrates — the platform, the arbitration body, or the court itself.
The demand question underneath: whether people accept a machine’s read of fairness as final, or only as a first pass before a person confirms it.
04 · The Agentic Delegation deskadjacent markets, moving now
Generative Governance
State staff now draft with a model.
Pennsylvania expanded generative AI to 3,000+ state employees.
StateScoop
Regulated Reality
Lawmakers race to fence what synthetic content can do.
30 US states now have election-deepfake laws.
Public Citizen
The Explainability Demand
Regulators want the reasoning shown, not just the output.
EU AI Act Article 50 mandates AI-content transparency labelling.
European Commission
05 · The Wire
01
CaseCraft AI launched a UK small-claims platform (2025).
Legal IT Insider ↗
02
AAA to roll out an AI arbitrator for construction disputes
Bloomberg Law ↗
03
Arbitral institutions add AI to case handling and awards
Faegre Drinker ↗
04
Abu Dhabi courts plan AI claims-review tools by year-end
Khaleej Times ↗
The panel on this signal
They agreed on
–The model enters as a triager and drafter, not yet a decider that can bind on its own.
–The moat is institutional accreditation, not the model underneath.
–Trust in the ruling, not processing speed, is what gates how far this spreads.
The regulatory analyst disagreed on
–Documented use cases are limited to small-claims and construction disputes; generalising from those narrow, procedurally simple domains to a broad shift in dispute resolution overstates what this evidence base actually supports.
Know more about the method →
✦How this signal was built
01 · Panel
A Delphi panel of 4 AI analysts, each reasoning independently.
→
02 · Debate
They challenge and revise each other’s reads over two rounds.
→
03 · Evidence
Every claim is tied to a cited, linked source on The Wire.
→
04 · Synthesis
A moderator writes it into one read, disagreement and all.