jhana

Science &
Technology

VerificationPreliminary results

We decouple taste (penumbra) from reading (core), and certify whether the minimum evidence was considered. That takes the liability out of full automation, and gives models a signal they can hill-climb.

Motivation

DD tasks are where the reading is: audit, forensics, scrutiny, and eDiscovery. Each is a large, closed document set of >>2M tokens, with no external research. The answer is in the file, for whoever reads all of it.

Problem

Open texture, no oracle: reasonable experts disagree on the same evidence. Only 44% of USSC judgments are unanimous! So there is no answer key to grade an agent against.

Bound

A high score cannot hide skipped evidence. A 99% score can hide only required evidence within 1% of the weight, even for an agent trained to maximize the score. The bound is proved, and tightCˆw(T)≥1−ε ⟹C∗(T)≥1−μw(ε) (tight)
μw(ε): largest fraction of required units with total weight ≤ε.

Verifier scores correlate with human lawyers better than naively reading more

Lawyer-marked evidence read 0%50%100%0%50%100% Reading volume τb = 0.368 0%50%100% Verifier score τb = 0.548
40 contract runs. Also beats BM25 via LLM query gen. 11,356-page map evaluation: AUC .963 vs .900 BM25.

Agents use 31⁠–⁠92% fewer tokens per coverage point while mean coverage improved by >21.5 pp

naive Harness n tokensper point .49.26.34.13.15 .68GPT 5.6 Luna23−31%.52GPT 5.4 Mini22−53%.46Gemini 3.5 Flash24−32%.42Gemini 3.1 Flash Lite24−41%.35Gemini 3.1 Pro24−92%

Convergence to coverage 1.0

Mean 35% → 98%; all tasks converge ≥95% in ≤8 rounds of repair.

024680.51 repair round stop at τ = .95 35% 98%

Not yet saturated by top models, well-differentiated signal

This verifier can be hill-climbed to finetune agents and harnesses for the law.

.30.35.40.45.50 naive control .345 verifier score GPT 5.6 Sol (high)GPT 5.6 SolGPT 5.6 TerraGPT 5.6 LunaOpus 5Sonnet 5Haiku 4.5DeepSeek V4 ProDeepSeek V4 FlashKimi K3Grok 4.6 .498.443.432.405.463.410.368.405.381.445.415
Real work data for evaluations

Real tasks from 16K+ lawyers and 650+ judges, and 300K+ disputes followed from filing through disposal. Build and evaluate agents against what actually happened in court, so they avoid liability, win more, and cost less.

2,440,503 requests by location

Through September 2026
15,121accounts
63,483agent tasks
6,999,784document pages processed
2,506towns and cities with active users
9,527minutes of audio recorded
774,566documents uploaded

Model‑ & harness‑agnostic e2e orchestration

Swarmmany agents work every step at once
End to endeach claim runs one line from arrival to close, like a factory floor
Sovereignthe whole stack can run on a local model in your cloud or on your premises, with no data leaving

Runs on

  • Any model

    A lab API, or a local model

  • Open weights

    Hill-climbed against the verifier to the Pareto frontier

  • Any harness

    Ours, or yours

  1. Capture & Connectors

    Call logs, ticket history, notices, data enhancement, and the CFPB clock

  2. Data Forensics

    Receipt vs. ledger, missing recordings, conflicting dates, discovery, productions, and deemed admissions

  3. Routing

    Arbitration clause, motion to compel, Rule 68, settlement authority, and trial

  1. Filings/Comms

    Answer and removal clocks, answers, early motions, e‑filing, service, and docketing

  2. Hearings & IRL Events

    AI transcripts and next dates

  3. Disposal & Follow-ups

    Release, dismissal with prejudice, tradeline correction, enforcement, closure, and root cause

Pattern informs prediction A shared payment-posting issue or cancellation flow

All 1,642Due this week 38Needs you 46

1,284closed by agents, flex counsel signed1,031 before suit · 88 dismissed · 165 settled under reserve
312with flex counsel for sign‑off
46need you
ClaimWhat the agent didOwnerNext actionDue
Payment disputeMatched the receipt to its posting time; drafted the answerFlex counselFile responseNov 18
RepossessionMatched the notice theory to 14 open claims; drafted the motionPanel firmMove to dismissNov 20
Credit reportingFound the reporting error; priced the fee clock; drafted the releaseYouEscalate to settlementNov 22

AI audit

AI safety &
human autonomy

Explainability, audit trail, and human liability.

Tokenomics
& ROI

(Re)Staffing, time & token attribution, and business development.

Sovereignty &
client privilege

Vendor exposure, data privacy, and client disclosures.

Forward-deployed technology

L0

Suit

Formalization & verification. Context & completeness.

L1

Para Agent

Model-agnostic or local harness with firm guardrails.

L2

Spatial

Role-based IT & cloud substrate for humans & AIs.

L3

Temporal

Assembly-line automation: triggers, schedules, and goals.

Start with a conversation.

Talk to our team