legal-os · leader view

Fault-Line Radar — the screen

The page a leader opens inside Legal AI OS.
One chart, three things to build now, and a record that grades every call.

The screen answers one question before you scroll: what should we build before the rules land?

See the corner. Build the corner.
The radar map — two lenses
0 0 2 2 4 4 6 6 8 8 10 10 threshold 7 BUILD NOW 1 2 3 4 5 6 7 8 9 10 11 Right = the law is closer to acting → Up = the market made the fix table stakes →
now soon later Dot size = queue score · number = queue rank · faint line = where it sat 90 days ago
For a firm: which control to build first. Right = the law is closer to acting (L2). Up = the market already made the fix table stakes (L3). The shaded corner, past 7 on both, is the build-now queue.
Every fault line — click to expand
1 Verification-by-default with an audit trailSo every AI-assisted filing carries proof it was checked, and certification is a byproduct. now 7·10·8
L1 capability6.7 (+1.2)
L2 ruling pressure9.8 (+1.8)
L3 adoption8.1 (+1.1)
Why now

Courts first punished fake AI citations after the fact (Mata v. Avianca was the first such sanction). Now more than 300 judges require lawyers to disclose AI use up front, and that patchwork of local orders is converging toward a single certification standard.

What to put in place

Verification-by-default with an auditable trail, so certification is a byproduct.

Model rules: 3.3, Rule 11

L2 ruling evidence
L3 adoption evidence
  • T3CNA adds AI-governance questionnaires to malpractice renewals · 2026-06-15
2 Documented verification + trace logsThe competence duty moves from a signature to a record of what was checked, how, and by whom. soon 7·10·8
L1 capability6.7 (+1.2)
L2 ruling pressure9.8 (+2.3)
L3 adoption8.1 (+1.6)
Why now

Signing a line that says 'a human reviewed it' is no longer enough. Courts and bars increasingly expect a documented verification process: a record of what was checked, how, and by whom.

What to put in place

Logging, trace analysis, evaluation records as the compliance artifact.

Model rules: 5.1, 5.3

L2 ruling evidence
L3 adoption evidence
  • T2Federal court excludes expert for AI-hallucinated citations · 2026-08-31
  • T2ABA Formal Opinion 512 — Generative AI Tools · 2024-07-29
3 Firm-wide training + governance programThe duty shifts from understanding one tool to running a program across all of them. soon 4·10·8
L1 capability4.5 (+2.0)
L2 ruling pressure9.6 (+3.6)
L3 adoption8.2 (+2.7)
Why now

The duty is shifting from 'understand the tool you use' to 'run a governance program' — training, written policies, and oversight — with mandatory AI continuing-education requirements spreading state by state.

What to put in place

Training and enablement at scale (iTrain bet; AI Build Lab proof).

Model rules: 1.1

L2 ruling evidence
L3 adoption evidence
  • T3Stanford: Westlaw AI 34% error rate, Lexis+ AI 17% · 2026-06-01
  • T2New Jersey requires technology CLE including AI — competence as a program requirement · 2025-04-01
  • T2ABA Formal Opinion 512 — Generative AI Tools · 2024-07-29
4 Governance as an insurable artifactCarriers now price malpractice on whether you can show a program, not just claim one. now 2·9·8
L1 capability2.5 (+0.0)
L2 ruling pressure8.7 (+0.2)
L3 adoption8.4 (+0.9)
Why now

Malpractice insurers are adding AI-governance questions to policy renewals and pricing coverage on the answers. They move faster than bar regulators, so insurance, not the rules, becomes the real gatekeeper for whether a firm can use AI at all.

What to put in place

The governance program as an insurable artifact: model inventory, approval process, oversight documentation, measurement.

Model rules: 1.1, 5.1, 5.3

L2 ruling evidence
L3 adoption evidence
  • T3CNA adds AI-governance questionnaires to malpractice renewals · 2026-06-15
5 One operating model, to the strictest standardDisclosure, competence, and confidentiality collapse into a single set of controls. now 2·9·8
L1 capability2.5 (+0.0)
L2 ruling pressure9.3 (+1.8)
L3 adoption7.7 (+0.7)
Why now

The rules are a patchwork — the EU AI Act, Colorado, Texas, 35-plus state bars, 300-plus judges — each slightly different. The patchwork itself is the pressure: a firm working across jurisdictions needs one operating model built to the strictest standard.

What to put in place

The unifying operating model. This is the whole thesis.

Model rules: 1.1, 1.6

L2 ruling evidence
L3 adoption evidence
  • T1EU AI Act obligations take effect · 2026-08-01
6 Data-flow mapping + vendor attestationWhat client data goes into a model, and where it lands, becomes an ethics question. now 4·9·7
L1 capability4.5 (+0.0)
L2 ruling pressure9.4 (+2.4)
L3 adoption7.0 (+1.0)
Why now

AI tools that learn from what you feed them, or mix data across matters, put client confidentiality at risk. Firms are being pushed to map where client data goes and get written no-training guarantees from vendors before a tool is approved.

What to put in place

Data-flow mapping and vendor attestation at tool onboarding.

Model rules: 1.6, 1.7, 1.9

L2 ruling evidence
L3 adoption evidence
  • T2ABA Formal Opinion 512 — Generative AI Tools · 2024-07-29
7 Tool-certification benchmark (NERVE)“It works” gives way to a measured score you can put in front of a client. soon 9·7·8
L1 capability8.6 (+2.1)
L2 ruling pressure7.2 (+1.2)
L3 adoption8.0 (+2.5)
Why now

Independent studies now publish how often each legal AI tool gets things wrong. Once error rates are public, choosing a tool becomes a measurable competence decision rather than a matter of trust.

What to put in place

You already built one: NERVE. A benchmark is a compliance instrument.

Model rules: 1.1

L2 ruling evidence
L3 adoption evidence
  • T3CNA adds AI-governance questionnaires to malpractice renewals · 2026-06-15
  • T3Stanford: Westlaw AI 34% error rate, Lexis+ AI 17% · 2026-06-01
8 Human-decides gates + scope-limitingAs agents act on their own, supervision has to be designed in, not assumed. soon 4·7·6
L1 capability4.0 (+0.0)
L2 ruling pressure7.1 (+0.6)
L3 adoption6.2 (+1.2)
Why now

AI is shifting from drafting text to taking actions on its own across multiple steps (so-called agents). 'Just review the final output' breaks down when no single person saw every step, reopening old duties around supervision and the unauthorized practice of law.

What to put in place

Human-decides gates and scope-limiting by design.

Model rules: 5.1, 5.3, 5.5

L2 ruling evidence
L3 adoption evidence
  • T3AI AGENT Act would require scope-limited, revocable, logged agent authorization · 2026-08-29
9 Value-delivered measurementBilling for hours a machine did in seconds becomes a question you have to answer. later 2·7·5
L1 capability2.5 (+0.0)
L2 ruling pressure6.8 (+2.3)
L3 adoption4.7 (+1.7)
Why now

AI collapses the hours a task used to take, which strains the billable-hour model and raises the question of what counts as a 'reasonable fee' when the work now takes minutes.

What to put in place

Measure value delivered, not hours saved.

Model rules: 1.5

L2 ruling evidence
L3 adoption evidence
  • T2ABA Formal Opinion 512 — Generative AI Tools · 2024-07-29
10 Guardrailed strategy sim (no actor prediction)Predicting judges runs into rules some courts have already started writing. soon 4·6·3
L1 capability4.0 (+0.0)
L2 ruling pressure6.4 (+1.4)
L3 adoption3.0 (+0.0)
Why now

Tools that predict how a specific, named judge will rule cross the line from analytics into profiling. France has already criminalized it, and US regulation is likely to move in the same direction.

What to put in place

Strategy stress-testing with human-reviewed, explicitly-uncertain priors; never sell 'predict the judge'. Persona models are low-confidence parameters, never determinative. (France already criminalized judicial profiling; US direction is toward rules.)

Model rules: 3.5, 8.4, —

L2 ruling evidence
L3 adoption evidence

Not scored yet — no market signal recorded.

11 Tool-provenance documentationWhen a tool causes the harm, the contract decides who owns it. Most don’t say. later 2·6·2
L1 capability2.5 (+0.0)
L2 ruling pressure5.8 (+2.3)
L3 adoption2.5 (+0.0)
Why now

Today the lawyer carries all the risk when an AI tool fails; there is no clear legal path to hold the vendor responsible. Pressure is building, through disclosure mandates and product-liability proposals, to shift some of that liability onto the toolmakers.

What to put in place

Track for a shift; document tool provenance now.

Model rules: —

L2 ruling evidence
L3 adoption evidence

Not scored yet — no market signal recorded.

Build now — force-ranked, top three
#1now

Verification-by-default with an audit trail

Courts now expect a documented check on every AI-assisted filing, not a signature after the fact.

L2 9.8 +1.8 · L3 8.1 +1.1
#2soon

A documented verification process

Attestation gives way to a recorded process: what was checked, how, and by whom.

L2 9.8 +2.3 · L3 8.1 +1.6
#3soon

A governance program, not a tool ban

The duty shifts from understanding one tool to running a program across all of them.

L2 9.6 +3.6 · L3 8.2 +2.7
Everything else stays in the ranked queue below the cards. Years-out items collapse to a single just watch line.
The advisory view — govern vs deploy

Each fault line now carries a fourth reading, Enablement (E): can the market actually supply the tooling to stand this control up. And a companion view turns the landscape into the two calls a firm actually faces, kept separate so a required control is never delayed by the economics of the AI it governs.

Govern · stand up the control?

Required of us, and can we meet it. The billable hour never blocks it. Stand up now · build capacity first · no mandate yet.

Deploy · put AI on the work?

Gated by pricing, whether the work is capturable, and market enablement E ≥ 6. Deploy now · fix pricing first · defer · watch.

The full model, with the two-firm worked example, is on how it works.

Why the radar earns trust
6 / 9

real events called 90 days early

The screen carries its own calibration record: every run is checked against what actually happened, misses left in. For the full per-order breakdown and how the scoring works, see how it works.