Pre-emptive AI Control Diagnostics

Find the Break Point.
Before It Finds You.

AI-influenced systems break on their own timelines. The failure arrives the day before the audit, or the moment a regulator asks for evidence of control that turns out not to exist. Luminary Diagnostics finds the failure point first.

Pre-emptive Resilience
"Where is this system going to break under pressure? What is that pressure? What is the sequence of events, and how do I stop it before the mess I am in becomes irretrievable?"
Answered by BreakPoint
AI Authority
"Is the authority my governance documents describe actually exercisable at the moment AI-influenced decisions are made, or is governance a story we tell that no longer matches operational reality?"
Answered by Signal Pulse → BreakPoint → Exposure Pulse
Book a Review Meeting Take the Free Exposure Check →
Immediate
Signal Pulse report on completion
8 days
To BreakPoint Report
8 sectors
With published intelligence briefings
The Three-Stage Approach

Triage. Diagnose. Monitor.

Three stages. Each produces a written determination. Each builds on the one before it. You can enter at any stage, but the journey is designed to be followed in sequence.

Stage 01: Triage
Signal Pulse

"Does authority over our AI-influenced decisions actually exist, or is governance a description of something that is no longer present?"

Signal Pulse Automated, £997
Stage 02: Diagnose
BreakPoint

"Where are the control gaps? What is Real, Assumed, Absent? Will our governance hold under pressure, and what does it need to feed our Target Operating Model?"

BreakPoint Expert-led.
Stage 03: Monitor
Exposure Pulse

"Is the control we established still holding, as AI use changes and agents take on more?"

Exposure Pulse Autonomous
Stage 01: Triage

Signal Pulse: a constitutional threshold test

Six questions across five dimensions. One determination: PROCEED, PAUSE, or STOP. Completed in 12 to 18 minutes by the person who carries accountability. The Constitutional Signal Report is delivered immediately on completion.

It establishes whether authority over your AI-influenced decisions is real, visible and exercisable at the point decisions are made, or only described. The right question before any deeper diagnostic begins.

Get Signal Pulse, £997 → Includes Signal Debrief, 3 months Foundation-level VC membership + more

£997 credited in full against BreakPoint.

Stage 01: Triage / £997 / Automated
"Is authority over our AI-influenced decisions real, visible and exercisable, or is governance a description of something that is no longer present?"
Six questions across decision ownership, authority drift, visibility and activation
Examines a specific AI-mediated process you name, personalised to your situation
PROCEED / PAUSE / STOP determination with full interpretation
Constitutional Signal Report generated immediately on completion
Completed in 12 to 18 minutes. No interviews. No waiting.
Automated: report delivered immediately on completion
Stage 02: Diagnose

BreakPoint: The AI Control Gap Assessment

One structured executive interview. Four questions. Every AI-influenced decision point in scope examined against three criteria. The result is not a score. It is a written determination of where human authority is Real, Assumed, or Absent.

Every decision point receives one of three determinations

●
Real
Human authority at this decision point is evidenced, documented and defensible. A named individual exercised genuine authority, not process oversight.
▲
Assumed
Human authority is claimed but cannot be evidenced from the organisation's own documentation. The assumption exists. The evidence does not.
✕
Absent
No identified human authority at this decision point. AI is determining outcomes with no evidenced human intervention between input and consequence.

The four questions BreakPoint examines

1
Where is this system going to break under pressure, and what is the consequence when it does?
2
What is that pressure, and when does it arrive?
3
What is the sequence of events that will cause it to happen?
4
How do I stop it, before damage is done, before a regulator finds out, before the board faces it unprepared?
Output
The BreakPoint Report
A board-level written determination mapping Real, Assumed and Absent human authority across the AI-influenced decision chain in scope. Built from your organisation's own evidence. Designed for board, regulator or insurer presentation.
One 90-minute structured executive interview
Written determination delivered within 8 working days
Real / Assumed / Absent mapped across the decision chain
Designed for board, regulator or insurer presentation
No framework imposed, built from your own evidence
Structured input to Target Operating Model redesign
Book a Review Meeting See a sample report →

We do not advise. We determine. If a regulator demands evidence of control, a maturity score will not satisfy them. A written constitutional determination, built from your own evidence, will. For organisations redesigning their Target Operating Model in the AI era, BreakPoint is the input TOM architects have been unable to source until now.

New governance risk: 2026

Governments can now unilaterally restrict access to foundational AI infrastructure across entire jurisdictions. If the AI systems at the centre of your decision chains become unavailable, the constitutional determination BreakPoint produces, built from your own evidence and not dependent on any AI system, is the only independent map of where human authority existed. That is what a regulator, insurer or acquirer will ask for. It is now also what your board needs before the risk crystallises.

BreakPoint in Practice

See what a BreakPoint assessment actually finds.

Three composite assessments across Insurance, Wealth Management and Asset Management, each constituted from published regulatory evidence. Real. Assumed. Absent.

View the assessments →
Sector Intelligence

What BreakPoint finds in your sector

Eight sector intelligence briefings, each built from the published regulatory evidence base for that industry. Each maps where human authority is most at risk and the regulatory pressure driving it. Predictive inferences from evidence, not assertions of findings.

Briefing 01
Insurance
AI underwrites risk. AI sets premiums. AI processes claims. In how many of those decisions can you name the human who held authority?
Read the briefing →
Briefing 02
Wealth Management
The complete absence of published override rate data in UK wealth management is itself a governance signal.
Read the briefing →
Briefing 03
Asset Management
A senior manager who certifies accountability for a process driven by AI does not hold accountability. They hold liability.
Read the briefing →
Briefing 04
Banking
84% of firms have a named accountable person for their AI framework. Framework accountability is not decision-level accountability.
Read the briefing →
Briefing 05
Defence
Human in the loop is the stated doctrine. Whether it describes operational reality, at decision level rather than policy level, is the question no published audit has yet answered.
Read the briefing →
Briefing 06
Pharma & Life Sciences
PV AI surfaces what it detects. What it does not detect does not reach human review. Nobody is named as responsible for what the AI misses.
Read the briefing →
Briefing 07
Retail
Retail AI makes ten million pricing decisions a day. No human makes any of them, but a human is named as accountable for all of them.
Read the briefing →
Briefing 08
Telecoms
AI manages network faults, sets pricing, detects fraud and routes customer outcomes. When a network decision affects emergency services, who is named as the accountable human?
Read the briefing →
Stage 03: Monitor

Exposure Pulse: Is control still holding?

Establishing control is not a one-time event. AI use changes, agents take on more and decision chains evolve. Exposure Pulse is autonomous continuous monitoring of whether the control position established through BreakPoint is still holding. It detects authority drift and control degradation before they become consequential.

Not a periodic review. A continuous governance layer, built for the agentic era.

Available from August 2026

The pace of change in agentic AI, and the emergence of new structural risk categories including infrastructural sovereignty, has prompted us to align Exposure Pulse with an entirely new constitutional framework before launch. The framework will be published ahead of availability. Pricing at launch.

Stage 03: Monitor / Autonomous Subscription
"Is the control we established still holding, and will we know within hours if it is not?"
Autonomous continuous monitoring of AI-influenced decision chains
Detection of authority drift and control boundary failures
Early warning before degradation becomes consequential
Subscription service, runs throughout the year
Designed for organisations operating agentic AI at scale
Autonomous: continuous visibility, not periodic review
Four Entry Points

The same methodology. Four doors into it.

Different buyers arrive at the AI control question from different directions. The diagnostic methodology is the same, but the urgency, the framing and the consequential question differ.

Door 01
Boards & Executive Teams

"Do we have governance standing if this goes wrong?"

Boards are first to be asked the question and last to have the evidence. BreakPoint gives the board a defensible written determination of exactly where human authority holds and where it does not, before a regulator, insurer or claimant asks first.

Door 02
Internal Assurance Teams

"Can we evidence control to an external auditor?"

Assurance teams need evidence, not frameworks. BreakPoint produces constitutional determinations from your own evidence, the kind that can be handed to an external auditor, regulator or insurer without qualification.

Door 03
Professional Services & Consultancies

"Can we deliver genuine AI governance assurance to our clients?"

Consultancies can introduce BreakPoint as the independent constitutional assessment their clients need. Structural separation is preserved by design. Luminary determines, the consultancy advises.

Door 04
Regulators & Supervisory Bodies

"Can this sector provide independent, evidenced assurance that human authority over AI is real, not assumed?"

Luminary Diagnostics operates as an independent third-party assessor. The sector intelligence briefings provide the evidenced baseline against which sector-level assessments are made.

The Closing Window

AGI has not arrived as an event. It is arriving as a condition, and its economic shadow is already inside your organisation.

Now: The Jagged Frontier
Superhuman in some domains. Near-zero in others.
The Stanford AI Index 2026 reports frontier systems matching or exceeding human performance at PhD-level science, competition mathematics and professional legal reasoning. The same systems score 0.51% on ARC-AGI-3, where humans score 100%. Extraordinary capability alongside elementary failure, distributed unevenly across your workflows right now.
Near Term: The Economic Shadow
Parts of the AGI definition are already met.
DeepMind's Levels of AGI treats capability as a graded spectrum, not a binary event. The economic shadow is present now. When broader capability consolidates, control either exists or it does not. There is no position from which to establish it retroactively.
Beyond: ASI
The question becomes existential.
The organisations without demonstrable human authority at this inflection point are not simply behind. They are structurally unable to catch up on their own terms.

Most governance frameworks were designed for ANI, narrow AI operating within defined parameters. We are past that.

The Stanford AI Index 2026 documents benchmark saturation. METR task-horizon data shows exponential growth in autonomous AI work. Frontier systems exceed expert human performance across multiple professional domains while remaining poorly understood by the organisations deploying them. That is the governance gap, not that AI is too powerful to control, but that most organisations have not diagnosed the shape of what they are governing.

Sources: International AI Safety Report 2026  ·  Stanford AI Index 2026  ·  DeepMind Levels of AGI  ·  METR task-horizon research  ·  ARC-AGI-3 technical report
Diagnose your position now →
Free Self-Assessment

See where you are most exposed, in 5 minutes

Five questions. No sign-up. The result tells you which diagnostic fits your situation and how urgently you need to act.

Question 1 of 5
If AI influenced a consequential decision last month, do you know which decisions?

Book a Briefing
Steve Butler, Luminary Diagnostics
Author of nine books on artificial intelligence and governance
Creator of the Six Laws of Epistemic Opposition
Creator of the Eleven Laws of Human Governability
Creator of the Seven Laws of Agentic Safety
Creator of the Laws of Pre-emptive Governability (LPG)
Creator of the Constitutional Agentic Operating System (CAOS)
Frameworks on NIST public record, June 2026
Co-founder, Luminary Diagnostics
The Intellectual Foundation

Built for the era that's arriving. Not the one that's passing.

Every existing AI governance framework was designed for narrow AI: compliance checklists, maturity models and risk registers built for a world where AI does one defined thing within bounded parameters. That world is already behind us.

Luminary Diagnostics was built on Steve Butler's original constitutional frameworks: the Six Laws of Epistemic Opposition, the Eleven Laws of Human Governability, the Seven Laws of Agentic Safety, and, established on the NIST public record in June 2026, the Laws of Pre-emptive Governability and the Constitutional Agentic Operating System. LPG identifies ten governance territories where human authority is structurally at risk as AI capability advances toward AGI and ASI. CAOS provides the constitutional operating architecture for agentic AI deployment. Neither is adapted from existing frameworks.

When Exposure Pulse launches in August 2026, it will be the first monitoring product in the world aligned to a constitutional framework purpose-built for the agentic era. Our competition governs yesterday's problem. Luminary Diagnostics is built for the one arriving now.

Steve Butler is the author of nine books on AI governance, constitutional intelligence and epistemic safety. His forthcoming tenth, The Closing Window: Human Authority in the Age of Agentic AI, is in development. His foundational frameworks are on the NIST public record.

Luminary Diagnostics and AITru share a common founder but operate as independent entities. Luminary diagnoses. AITru governs. Structural separation is maintained to preserve diagnostic independence.

Common Questions

What people ask before they book

Where do I start: Signal Pulse or BreakPoint?+
If AI is present in your consequential decisions and you need to understand the control gaps, go directly to BreakPoint. If the territory is still unclear, start with Signal Pulse, it establishes the factual baseline in 12 to 18 minutes before any expert-led diagnostic begins. Most organisations with a specific system or process at risk go straight to BreakPoint.
Can a BreakPoint assessment feed into our Target Operating Model?+
Yes. BreakPoint produces a precise map of where human authority is Real, Assumed or Absent across each consequential decision point. For TOM architects designing around AI-influenced decision chains, this is the structured constitutional input that has not previously been available, and which maturity scores cannot provide.
What is Exposure Pulse and how does it work as a subscription?+
Exposure Pulse is the Monitor stage, autonomous continuous monitoring of whether the control position established through BreakPoint is still holding. It detects authority drift and control degradation before they become consequential, running throughout the year rather than as a periodic review. Subscription terms are discussed during the initial briefing.
What makes this different from a governance audit or maturity assessment?+
Luminary Diagnostics makes constitutional determinations: binary answers to whether human authority over AI-influenced decisions exists at each decision point examined, drawn from your own evidence. A governance audit scores maturity against a framework. If a regulator or insurer demands proof of control, a maturity score will not satisfy them. A written constitutional determination will.
How long does each diagnostic take?+
Signal Pulse: completed in 12 to 18 minutes, report delivered immediately on completion. BreakPoint: a 90-minute executive interview, written report within 8 working days. Exposure Pulse: runs continuously as a subscription from setup.
Is Luminary Diagnostics connected to AITru?+
Luminary Diagnostics and AITru share a common founder but operate as independent entities. Luminary diagnoses. AITru governs. Structural separation is maintained deliberately to preserve diagnostic independence.

Ready to find out where you stand?

Every engagement starts with a review meeting. No preparation required. Thirty minutes.

The First Step: Always
Book a BreakPoint Review Meeting
A structured 30-minute conversation to map your AI control exposure, identify which decision chains are most at risk, and establish whether BreakPoint is the right next move. You leave with a clear picture of where you stand. No obligation, no sales process.
Book the Review Meeting
Not Sure Yet
Take the Exposure Check First
Five questions. Five minutes. No sign-up. Find out where you are most exposed before you book anything.
Take the Check →
Already Confirmed
Referred by a Partner?
If you have already been briefed on BreakPoint through a consultancy or introducer, book directly. The review meeting is still the starting point, your introducer has already done the framing.
Book Directly