Prompt Injection Is a Governance Failure Wearing a Technical Costume.

Posted on by

Every prompt injection headline reads like a technical failure. A model got tricked. A filter didn’t catch it. The vendor needs to patch something.

That framing is comfortable, and it is wrong. The technical trick is real. The governance failure sitting underneath it is the actual story, and it is the one almost nobody in the room wants to own.

 

Why the Trick Works in the First Place

The mechanism is architectural, not a bug in the usual sense. Large language models treat the system prompt, the user’s request, and any text retrieved from an external source as a single stream of tokens. There is no reliable internal boundary between an instruction and a piece of data. A hostile sentence buried in a document, a web page or a support ticket can carry the same authority as a command typed directly by a trusted user, because the model was never built to tell the difference.

OWASP’s 2026 State of Agentic AI Security and Governance report found prompt injection now maps to six of its ten top categories for agentic applications, up from a mostly theoretical concern in the 2025 edition to one backed by documented breaches and tracked vulnerabilities. Coding agents dominate the attack data, and only 37% of organisations report having a policy in place to even detect unauthorised AI deployments running inside their own environment.

 

The Failure Is a Control Boundary, Not a Model Flaw

This is where the governance framing actually matters. Prompt injection succeeds because enterprise workflows assume the model can reliably tell trusted instruction apart from hostile text, an assumption that fails the moment one interface carries user intent, retrieved content and tool-facing control signals in the same session. Most organisations respond by treating guardrails as a static filter list rather than a real system of content separation, monitoring and authorisation. A filter can catch a known bad phrase. It cannot answer the actual governance question, which is who controls what the system is allowed to do once it has been steered.

Security researcher Simon Willison’s “lethal trifecta” names the actual risk condition plainly: an AI agent with access to private data, exposure to untrusted content, and the ability to communicate externally, all three at once, is where exfiltration happens. Meta’s own internal guidance treats those three properties as a budget rather than a checklist. Combine all three and the agent needs a human in the loop before it acts, not after.

 

Why This Keeps Getting Treated as IT’s Problem Alone

Handing this to the security team as a patching exercise misses what the data is actually showing. A model update might close one exploit path. It will not answer who approved an agent’s access to a customer database, why that same agent can also send emails externally, or what happens the day it does both at once because nobody ever wrote down that it should not be allowed to. Those are ownership questions, not model questions, and ownership questions do not get solved by a vendor release note.

 

What Governance-First Actually Requires

Start by classifying every channel an agent reads from according to trust level, and keep untrusted content out of instruction scope entirely rather than hoping the model sorts it out at runtime. Quarantine tool access behind explicit policy gates, so an agent combining private data access, untrusted content and external communication needs sign-off before it can act, not a retrospective audit after it already has. Treat a pattern of near-miss prompts as an abuse signal worth escalating, not a string of isolated one-off incidents each closed out individually.

All of it is the same governance discipline organisations already apply to identity and access management, pointed at a new kind of actor that happens to run on language instead of credentials, not a new technology purchase.

 

Who Approved This, and Did They Know What They Were Approving

Before the next prompt injection incident gets logged as a technical exploit, ask who actually approved the access the exploit relied on.

If nobody can answer that cleanly, the model was never the vulnerability. The governance around it was.

Your Twenties Reward Hustle. Your Forties Reward Judgement. Few People Notice the Shift in Time.

Posted on by

Nobody tells you when the game changes. You keep playing the twenties game in your forties and quietly wonder why the same effort stopped yielding the same results.

The shift is real, not a mood. It has a name in the research and a shape most people never get to see clearly enough to plan around.

 

What Actually Peaks in Your Twenties

Stanford’s Center on Longevity has the cleanest explanation of why hustle works so well early. Fluid intelligence, the raw ability to solve new problems quickly without relying on prior knowledge, peaks as people enter their third decade. That is a genuine cognitive advantage, not a myth about youthful energy. Speed, pattern-spotting on unfamiliar problems, and the appetite to grind through volume are all things a twenty-something brain does better than it ever will again.

That is exactly why hustle gets rewarded so visibly early on. It is the highest-value thing on offer at that stage, and organisations are right to reward it.

 

What Actually Peaks Later, and Why Nobody Notices

The mistake is assuming that advantage holds. It does not, and something else takes its place instead of just fading. Crystallized intelligence, the accumulated knowledge and judgement built from lived experience, keeps rising well into the seventh decade. Emotional intelligence peaks in the forties. Moral reasoning keeps improving through adulthood. Stanford’s own framing of it is the clearest version I have read: the twenty-five-year-old brings speed and fresh perspective, the fifty-year-old brings integration and judgement, the seventy-five-year-old brings wisdom and pattern recognition across decades none of the others have lived through.

Almost nobody plans their career around that curve, because almost nobody is shown it. The result is a lot of very capable forty-somethings still competing on twenties metrics, wondering why the hours are not converting into the same visible wins they used to.

 

Early Excellence Does Not Predict What You Think It Does

A December 2025 review published in Science, covering nearly 35,000 world-class performers across sport, music, chess and the sciences, found that early standouts and eventual world-class performers are largely different people. Peaking early does not reliably predict who ends up at the top later. The performers who lasted tended to explore broadly before specialising, building a wider base of judgement to draw on rather than narrowing down early and grinding one lane harder than everyone else.

The same pattern shows up at executive level. Generalist chief executives, the ones with genuine cross-functional experience rather than a single specialist lane, file more than double the patents annually compared with specialist peers, with 55% higher originality ratings. Jeff Bezos building AWS and Satya Nadella’s Microsoft turnaround both drew on lateral experience outside their original speciality, not deeper hustle inside it.

 

What Judgement Actually Looks Like at Work

Judgement rarely looks impressive in the moment, which is part of why it goes unrewarded for so long. It looks like not escalating something that will resolve itself on its own, and letting a good-enough answer stand instead of spending three more hours perfecting a version nobody asked for. It looks like knowing which fight is worth having this quarter and which one can wait, a decision hustle never had to make because hustle just took every fight.

The forty-something who is still measuring their own value in hours and visible output is applying a twenties scorecard to a role that has already moved past it. The actual multiplier at that stage is rarely personal execution. It is knowing which three things matter this month and being willing to let the other twelve go undone.

 

The Shift Worth Naming Out Loud

Nobody sends a memo when hustle stops being the highest-value thing you offer and judgement takes over. It happens quietly, somewhere in the decade nobody warns you about, and the people who notice early enough to adjust are the ones whose forties look like a promotion instead of a plateau.

The question worth asking yourself this year is not whether you are working hard enough. It is whether you are still being rewarded for the twenties game, or you have quietly moved into a different one without updating your scorecard.

Pre-Mortem: The Accountability Question the Mills Review Left Open

Posted on by

On 6 July 2026, the Financial Conduct Authority published the Mills Review, its examination of how AI will reshape retail financial services in the UK. The review covers seven recommendations across the regulatory perimeter, oversight architecture, and the transition to autonomous decision-making. It names the accountability gap at the centre of autonomous AI trading. It does not close it.

This is the thirteenth piece in the Pre-Mortem series. Five questions, applied to the public record, before the outcome is known.

 

The Bet

UK firms deploying autonomous trading AI are betting that the Senior Managers and Certification Regime (SMCR), the framework that holds named executives personally accountable for conduct failures in their area of responsibility, covers their position through general senior manager oversight. The FCA has been clear that delegating a decision to an algorithm does not transfer senior manager liability to the algorithm. The bet is that this principle, correctly stated and on the public record, can be demonstrated in practice before an enforcement case defines what demonstrating it actually requires.

 

The Assumption

Seven recommendations. One question still without an answer:

When an autonomous trading system executes a decision at machine speed, without pausing for human approval of the individual trade, which specific senior manager function is accountable if that decision causes a customer loss, and what does demonstrating adequate oversight of a system like that actually require?

The Mills Review acknowledged the problem directly. Without guidance, the review found, the combination of greater opacity in AI-mediated decisions and factors such as model drift makes it harder for the regulator to identify a de facto responsible individual, or for senior managers to evidence meaningful human control. Stakeholder feedback throughout the review called for clearer guidance on what constitutes the “reasonable steps” expected of senior managers. The review recommends the FCA develop it. The FCA has not yet published it. Every firm currently deploying autonomous trading AI is operating on the assumption that its existing accountability structure covers the gap. That assumption has not been tested in an enforcement case.

 

The Sequence

December 2019. SMCR extended to all FCA solo-regulated firms, completing its rollout across financial services.

27 January 2026. The FCA launched the Mills Review, acknowledging that AI in retail financial services had developed faster than the regulatory frameworks designed to govern it.

24 February 2026. Call for input closed.

6 July 2026. The review published seven recommendations. The FCA committed to adapting its regulatory frameworks as the transition to autonomous models continues. No guidance named a specific senior manager function as accountable for autonomous trading decisions. No guidance defined what “reasonable steps” requires for a system executing at machine speed without human review of individual decisions.

The capability reached the market before SMCR was tested against it. The review arrived after the capability. The guidance has not arrived yet.

 

The Pager

The FCA has confirmed there will be no dedicated Senior Manager Function for AI, and that accountability falls on existing functions. That is a clear policy position and it deserves credit for being stated plainly. The Treasury Select Committee has urged the FCA to publish guidance specifying the level of assurance expected of senior managers for AI-related harm. The Mills Review carried that request forward into its recommendations. The harder question is the one seven recommendations did not answer: when an autonomous trading system causes a customer loss, which specific function holder carries the call?

 

The Proof

There are no enforcement cases. The first case will establish what “reasonable steps” means in an AI trading context. The Mills Review is a process measure: it produced recommendations. The outcome measure worth watching is whether the FCA’s follow-on guidance names a specific function and defines the oversight standard in operational terms rather than principles alone. A principle restated is not a gap closed.

 

Verdict

If the FCA’s follow-on guidance names the senior manager function accountable for autonomous trading AI and defines what “reasonable steps” requires at the operational level, UK financial services will have resolved an accountability gap that every other major jurisdiction is still navigating. The review’s existence, the named individual who led it, and the seven published recommendations are genuine evidence that the FCA identified the problem and moved on it. Without operational guidance, the gap stays open. The first enforcement case will write the rule in the least comfortable setting available. That is a considerably worse way to write it.

Why Traditional Project Management Is Failing Modern Teams

Posted on by

Why Traditional Project Management Is Failing Modern Teams

Most project failures get blamed on execution. A missed deadline. A stakeholder who went quiet at the wrong moment. A scope that crept until nobody could point to when it happened.

Look earlier and the failure was already built in before a single sprint started.

A Forbes Technology Council analysis makes the point directly: misalignment gets embedded into the foundation long before execution begins, not manufactured somewhere in the middle. The team that won the deal rarely stays involved in delivery. The customer’s actual operating mindset only reveals itself once work is already moving. The incentives written into the contract often point delivery and client in different directions before day one. Teams execute a plan that was already structurally unsound before the first sprint started.

Traditional project management was never built to catch that kind of problem. Waterfall assumes you can define requirements fully upfront, lock them, and deliver against a fixed spec months later. That assumption survives about as long as the first change request. Priorities that shift inside a six-week planning cycle, which describes most programmes now, make a locked spec obsolete before it has even shipped.

 

Rigidity Is the Symptom. Something Else Is the Disease.

Here’s the twist most framework debates miss. Methodology by itself isn’t what separates the teams that deliver from the ones that don’t. PMI’s most recent Pulse of the Profession research found project performance sits at roughly 73.8% whether a team runs predictive, hybrid, or agile delivery, and whether people work remote, hybrid, or in-person. What actually moved the needle was business acumen: professionals strong in it posted 27% lower failure rates, regardless of which framework sat on the wall.

Framework still matters. It was just never the whole story, and treating “which methodology” as the central question misses where most programmes actually break.

A PM Solutions case study makes the same point at a larger scale. A U.S. staffing company with more than 8,000 internal and 90,000 contract employees had already tried and failed to stand up a PMO once. On the second attempt, the team built a hybrid methodology suited to the client’s actual environment, blending traditional project management, agile, and the touchpoints where each meets software delivery, then added a governance structure, portfolio visibility and resource planning on top of it. Within six months, every project flagged red under the new reporting system, more than $13 million worth of work, was recovered. One severely troubled multi-year project, over budget and behind schedule, was turned around in four weeks once it had a dedicated programme manager working inside that structure. Not a framework doctrine. Governance and visibility.

 

The Real Adoption Curve

That pattern shows up in the adoption numbers too. Hybrid delivery has grown 57% since 2020, while purely predictive approaches have fallen 24% over the past three years. The pattern behind that shift is simple: pure Waterfall and pure Agile were both answering questions the actual work wasn’t asking, and organisations are admitting it with their adoption numbers rather than in a strategy memo.

Rigid, one-size-fits-all implementation is what actually ages badly, more than the framework choice underneath it. Portfolios that mix delivery models by what the work actually demands, a predictable cadence for mature products, flow-based delivery for continuous work, fast validation cycles for early bets, consistently outperform portfolios that force every initiative through the same certified process regardless of fit.

 

What This Means for the Programme You’re Running Now

The practical shift isn’t abandoning structure for chaos. It’s building governance that travels with whichever delivery model fits the work, instead of assuming the delivery model is the governance.

Start by naming decision rights before the kickoff, not after the first dispute breaks something. Someone owns scope changes. Someone owns the call when a dependency slips. Everyone on the programme should be able to name both without asking. Build the escalation path into the plan itself rather than inventing one under pressure in week eight, and match the delivery model to the type of work in front of you rather than to what worked on the last programme. A regulated, multi-vendor transformation and a ten-person product team shipping a new feature are not the same problem, and forcing them through the same framework produces the same failure pattern twice.

Track what the framework was supposed to deliver, not whether the ceremonies happened. A team that ran every stand-up and still shipped nothing useful followed the process and failed anyway.

 

The Question Every Kickoff Should Answer First

Before the next programme gets a charter and a framework stamped on the cover page, ask a harder question first: who owns the decision when priorities collide, and does that answer exist in writing before the first sprint starts?

If it doesn’t, the framework on the cover page was never going to save the programme underneath it.

The $5.5 Trillion Bill for Doing Nothing About AI Skills

Posted on by

IDC put a number on the cost of enterprises not knowing what to do about AI skills: $5.5 trillion. Not by 2030. By the end of 2026.

That is not a distant workforce-planning problem. It is the price tag on decisions organisations are making, or more often deferring, this quarter.

The number comes from IDC’s survey of enterprise IT leaders across the US and Canada. It measures a specific kind of pain: product delays, lost competitiveness, and business walking out the door because the people needed to build and run AI capability were not there when the work needed them.

An IDC Spotlight Paper distributed via workforce-skills platform Workera applies the same figure directly to the AI skills conversation, and the application largely holds up. Most of what IDC’s original survey describes as a broader tech talent shortage is, in practice, an AI capability shortage wearing a wider label.

Three other figures make the same point from different angles. The World Economic Forum’s Future of Jobs Report 2025, surveying over a thousand employers and 14 million workers across 55 economies, found that 59% of the global workforce will need reskilling or upskilling by 2030, and that 11% of that group are unlikely to receive it. Indeed’s hiring data shows the share of job postings with “AI” in the title has more than tripled since 2022, from 2.6% to 8.3%. And PwC’s 2026 Global AI Jobs Barometer, built on more than a billion job advertisements, found that AI-skilled workers now command a 62% wage premium over comparable peers, up from 57% the year before and roughly 25% two report cycles before that.

These numbers describe a market that has already repriced itself, not a future state, while most enterprise workforce plans are still budgeted as though it hasn’t.

 

The Premium Is the Market Telling You Something

Wages move slowly almost everywhere except where genuine scarcity exists. A skill premium that has climbed from roughly 25% to 62% across three consecutive PwC survey cycles is not a normal labour-market signal. When a specific skill commands 62% more pay than the equivalent role without it, and that gap is still widening year over year, that is the market pricing in a shortage faster than most HR functions can respond to it, not a talent management curiosity.

Second Talent’s 2026 research puts a shape on that shortage: roughly 1.6 million open AI-related roles globally against around 518,000 candidates qualified to fill them, a demand-to-supply ratio a little over three to one. Every organisation competing for AI capability right now is competing inside that gap, and every quarter spent treating reskilling as a training-budget line item rather than a capital allocation decision is a quarter spent losing that competition to whoever moved first.

 

Why This Belongs in the Risk Register, Not the Learning and Development Plan

Most organisations still route AI reskilling through the same governance as any other training initiative: an L&D budget line, a completion metric, a once-a-year review. That treatment made sense when the skills in question were incremental. It does not hold up against a $5.5 trillion cost estimate and a wage market moving by double digits year over year.

A capital risk gets tracked differently to a training initiative. It gets a named owner, a quantified exposure, and a review cadence tied to the business calendar rather than the HR calendar. Few organisations apply that discipline to AI skills, because the function historically responsible for skills, HR, was never built to run risk registers, and the function that runs risk registers, finance and the PMO, was never asked to own workforce capability.

That gap in ownership is the reason the $5.5 trillion figure keeps compounding instead of shrinking, not a technicality.

 

What Actually Changes the Trajectory

Closing this gap requires three specific shifts most organisations have not made, not a bigger training budget.

Put a named executive owner on AI capability risk, distinct from whoever owns general L&D, with the same reporting rigour as any other material risk on the register. Measure the capability gap in the terms the wage market already uses: roles you cannot fill, roles you are overpaying to fill, and work you are declining because you lack the people to do it, not completion percentages on a training platform. And treat the reskilling decision as time-sensitive capital allocation, where every quarter of delay is a quarter in which the 62% premium, and the competitors already paying it, get further ahead.

 

The Window Is a Cost Curve, Not a Deadline

There is no single date after which the AI skills gap becomes unrecoverable. What exists instead is a cost curve that gets steeper the longer it is ignored, priced daily by a labour market that has already decided what AI capability is worth.

The organisations that treat this as a 2027 problem will be paying 2026 prices for it well into the decade. The ones already moving are the ones setting the price.

The Governance Training Nobody Budgets For

Posted on by

Nobody has ever failed a project because they didn’t know how to build a Gantt chart. However some failed because nobody taught them who was allowed to say no.

I have sat in quite a number of programme inductions, and every one of them covers the same ground. Scheduling. Budgeting. Risk logs. RAID templates. Reporting cadence. Then, somewhere around the second afternoon, someone puts up a slide about governance and the room’s attention visibly leaves the building. It is treated as the compliance module, the thing you sit through before you get to the real work. Nobody walks out able to say, with any confidence, who actually owns the call when a decision does not fit neatly on a template.

That gap is not an oversight. It is a choice organisations keep making, year after year, without ever naming it as one.

 

The curriculum has a hole in it

Ask a newly promoted programme manager to explain their RAID log and they will do it fluently. Ask them who has the authority to accept a risk above a certain threshold without escalating it, and watch the pause. That pause is the sound of someone realising they were never actually taught the answer, only ever expected to absorb it by watching more senior people long enough.

A 2026 AI Governance Gap Report surveying more than 500 HR professionals found that only 45 per cent of organisations provide AI literacy training to all employees, one concrete, measurable instance of the broader governance-literacy gap this piece is about. Two-thirds of HR teams are already using AI to shape compliance and policy decisions, and the same report found fewer than half have given employees even that AI-specific literacy training to keep pace. The gap shows up well beyond HR too, in every function that has ever built a decision-rights framework, filed it in a folder, and assumed the document itself did the teaching.

Documents do not teach. People do, and usually only the ones who were already senior enough to have picked it up somewhere else.

 

Decision rights are treated like folklore

Most organisations do not lack a governance framework. They have one, usually a good one, sitting in a policy library that almost nobody outside the PMO has opened. What they lack is a mechanism for turning that document into instinct.

The result is a workforce that learns decision rights the hard way: by guessing wrong in front of a steering committee, by escalating something trivial and being quietly told off for wasting everyone’s time, or by not escalating something serious and finding out only when it has become a crisis. Every one of those is an expensive way to teach a lesson that could have been taught in an afternoon.

This is where the “knowing-doing” research cited by Harvard Business Review becomes uncomfortable reading for anyone who runs a training budget. Two out of three managers say they are still uncomfortable having accountability conversations with their own people, despite most of them having sat through the leadership training designed to prepare them for exactly that. The problem was never a shortage of content. Knowing a framework exists and being able to act on it under pressure are two entirely different skills, and organisations keep training the first while assuming it produces the second.

Governance training suffers from the same fault line. Knowing there is an escalation policy is not the same as recognising, in the middle of a stressful Tuesday, that the decision in front of you is the one the policy was written for.

 

The training everyone skips because it looks obvious

There is a reason this particular gap survives budget reviews when almost nothing else does. Governance training looks like it should be simple, so nobody prioritises building it properly. Everyone assumes the framework document is self-explanatory, right up until the moment someone makes the wrong call and the post-incident review discovers that three different people had three different understandings of who was supposed to decide.

I have run those reviews. The finding is almost never “the framework was wrong.” Almost always, nobody had ever been walked through what the framework meant in a live situation, so everyone applied their own version of common sense, and common sense is not actually common.

 

What actually needs teaching

A longer policy document will not fix this. Longer documents get read less, not more. The fix is teaching people to recognise a decision point before it arrives, not after.

That means running people through real scenarios instead of abstract categories, trading “what is your escalation threshold” for “here is a supplier problem that looks small and is not, what do you do in the next ten minutes.” It means naming, out loud and often, the handful of decisions in your organisation that carry disproportionate weight, so people learn to feel the shape of one before it is labelled for them. And it means treating governance literacy the way you would treat safety training: refreshed, tested, and taken seriously enough that senior leaders visibly participate in it themselves, rather than left as a one-off induction module.

The organisations that get this right build fewer, better decision-makers rather than thicker governance frameworks: people at every level who can spot a genuine decision point on instinct, the same way an experienced engineer can hear an engine fault before the dashboard lights up.

You cannot budget for the crisis a decision creates and then refuse to budget for teaching people to see it coming.

 

 

Small Talk Is Not Wasted Time. It Is the Only Rehearsal for Big Trust.

Posted on by

Most executives treat small talk as the tax you pay before the real conversation starts. Research on negotiation and workplace behaviour suggests it’s closer to the opposite: the low-stakes rehearsal that determines whether the real conversation goes anywhere at all.

 

What the Research Actually Shows

A frequently repeated claim holds that people who make small talk before negotiating are four times more likely to reach agreement. That number doesn’t survive a check against the study it’s supposedly drawn from. The actual 2002 research behind it, published in Group Dynamics, found something more modest but still real: negotiators who “schmoozed” beforehand reported significantly higher rapport than those who didn’t, and reached an impasse less often, 40.6% of the time versus 60.7%, though that gap was only marginally significant. The finding itself is that small talk measurably raises rapport and modestly improves outcomes. It is not some four-times multiplier, and repeating the inflated number would undercut exactly the kind of precision this argument needs to be taken seriously.

 

The Mechanism, Confirmed More Recently

A 2021 study in the Academy of Management Journal tracked 100 employees across 978 daily workplace observations over three weeks and found small talk works through a specific, two-sided mechanism: it “enhanced employees’ daily positive social emotions at work,” which increased helpful, cooperative behaviour toward colleagues, while simultaneously disrupting people’s ability to concentrate on their actual tasks in the moment. Both things are true at once. Small talk builds the social capital that makes cooperation possible later, and it costs a small amount of focus right now. A 2024 qualitative study of 35 B2B professionals found the same rapport-building mechanism operating specifically in negotiation contexts, identifying genuine curiosity, active listening, and respect for boundaries as the actual ingredients, not just friendly chatter for its own sake.

 

The Counterargument Worth Taking Seriously

Not everyone in this field agrees, and the disagreement is worth taking seriously rather than editing out. Kim Scott, whose Radical Candor framework has shaped how a generation of executives think about direct feedback, has argued the opposite case directly: real trust with employees comes from substantive one-on-ones and working relationships, not casual chat, and treating small talk as the relationship-building mechanism risks substituting a comfortable habit for the harder work of actually knowing someone. That critique lands hardest in ongoing management relationships. The negotiation and cross-cultural research above is mostly about a different situation: the first few minutes with someone you don’t yet have a working relationship with, where there’s no substantive history to draw on yet, and small talk is the only tool available to establish enough trust for the real conversation to start at all.

 

Why This Matters More in Some Rooms Than Others

Research on Arab business negotiators found relationship-building carries even more weight in that context, with negotiators leaning on personal networks and trust-building as central to how deals actually get made, rather than an optional warm-up. The mechanism isn’t unique to any one culture. It’s just more visibly load-bearing in markets where trust is built through recurring personal contact rather than through contracts alone.

 

What This Means in Practice

Skipping small talk to “get to the point faster” isn’t efficient. It’s removing the only low-stakes moment where two people calibrate whether they trust each other, before the stakes get high enough that a miscalibration actually costs something. The application is genuine curiosity about the person in front of you, delivered before you need anything from them, rather than performed friendliness, so that when you do need something, the trust required to ask for it is already there.

Pre-Mortem: The Liability Chain Medicare’s AI Prior Auth Model Has Not Drawn

Posted on by

On 1 January 2026, the Centers for Medicare and Medicaid Services in USA launched the WISeR model in six states, introducing prior authorisation to procedures that traditional Medicare had always provided without it. Contracted companies now assess medical necessity using AI. Human clinicians are required to sign off on any denial. The Senate voted 46-50 in July 2026 to keep the programme running. One question has not been answered.

This is the twelfth piece in the Pre-Mortem series. Five questions, applied to the public record, before a programme has had the chance to succeed or fail.

 

The Bet

CMS is wagering that AI-assisted prior authorisation reduces unnecessary Medicare spend without producing the patient-safety incident that forces a political reversal. If WISeR delivers measurable waste reduction without a documented causal chain from AI denial to patient harm, it becomes the template for prior authorisation across Medicare nationally. If it produces that chain, a documented line from AI recommendation to denial to patient harm, it does not just end WISeR. It becomes the reference point that makes AI prior auth politically untouchable in federal health programmes for a generation.

 

The Assumption

CMS has answered every operational question about WISeR except this one:

When an AI recommendation leads a contracted clinician to deny care and a patient is harmed as a result, where does liability sit?

The model design places a human clinician between the AI output and the denial decision. That establishes a paper trail. It does not establish a liability framework. Contractors earn between 10 and 20 per cent of the savings generated by denials and lose that payment when a denial is overturned on appeal. That is a commercial penalty, not a clinical one. The Federal Tort Claims Act does not cover contracted entities. No federal court has tested whether a contracted clinician reviewing AI recommendations at volume carries the same duty of care as a treating physician making an independent clinical judgement.

The assumption doing all the work in this model is that the human review layer is accountability enough. That assumption has not been tested.

 

The Sequence

1 July 2025. CMS published the WISeR notice in the Federal Register and did not submit it to Congress under the Congressional Review Act. That omission would matter later.

1 January 2026. WISeR launched in New Jersey, Ohio, Oklahoma, Texas, Arizona, and Washington.

17 March 2026. The Washington Post published an exclusive: Medicare’s new AI gatekeeper was delaying care for seniors. The University of Washington’s medical system had nearly 100 patients waiting for epidural injections. In Arizona, Phoenix pain specialist Dr Matthew Crooks told Medscape that every epidural injection submitted in the first three months had been denied and described the system as completely nonfunctional and unsustainable. In Texas, initial AI approval rates ran at 62 per cent, against a 92 per cent national approval rate across Medicare Advantage.

25 March 2026. The Electronic Frontier Foundation filed a FOIA lawsuit against CMS in federal court in California, seeking records on WISeR’s AI algorithms, training data, bias safeguards, and the financial incentives paid to contractors. The suit confirmed that CMS had not made its AI methodology or vendor compensation structure publicly available seven weeks after launch.

6 April 2026. CMS published a Federal Register notice delaying prior authorisation implementation for certain services within the model to allow additional time for operational readiness. CMS also issued a corrective action order against one of its AI contractors. Both confirmed that the model’s operational design had not performed as intended in the first quarter.

12 May 2026. The Government Accountability Office issued its determination: WISeR met the Administrative Procedure Act definition of a rule and was subject to the Congressional Review Act. CMS had not made the required submission to Congress before the model took effect.

20 May 2026. Senator Ron Wyden and Representatives Suzan DelBene and Greg Landsman introduced resolutions of disapproval in both chambers, seeking to repeal WISeR under the CRA.

6 July 2026. Gold carding launched in Washington state. Providers achieving a 90 per cent affirmation rate across a minimum of ten prior authorisation requests become exempt from further review for covered services. Quarterly rollout to the remaining five states is planned.

16 July 2026. The Senate voted 46-50 against advancing the disapproval resolution. Party line. WISeR survived. The liability question the GAO had exposed survived with it.

The Pager

Dr Mehmet Oz, Administrator of the Centers for Medicare and Medicaid Services.

The message: WISeR’s accountability chain has not been drawn. The model places a contracted clinician between an AI denial recommendation and a Medicare beneficiary, but no published document establishes where negligence sits when a patient is harmed following an AI-assisted denial. The Federal Tort Claims Act does not cover contractors. Contractors point to the human clinician. Clinicians are reviewing AI output under volume pressure with no published duty-of-care standard for that specific context. When the first federal lawsuit tests this configuration, and one will, CMS will need a published framework, not a contract clause. That framework is easier to write before litigation than after.

 

The Proof

Gold carding is the model’s self-correction mechanism. If quarterly rollout reaches all six states and the 90 per cent affirmation threshold functions as a genuine quality signal, the AI layer contracts over time as trust is established. Proven providers exit prior auth. New entrants face the review. The model becomes calibrated rather than blanket.

If gold carding stalls or rollout criteria are applied inconsistently across jurisdictions, the AI layer expands without a release valve. Prior auth burden accumulates regardless of provider track record. The model becomes a cost-reduction instrument with no exit for providers who have earned one.

The proof of the bet is not the aggregate savings figure. It is whether WISeR, by the end of 2026, has published a liability framework and delivered gold carding in all six states. Without both, the model is running on the same untested assumption it started with.

 

Verdict

If CMS publishes a liability framework for AI-assisted denials before a federal case forces the question, and gold carding delivers consistent rollout across all six states, WISeR will be the strongest government evidence yet that AI-assisted utilisation review can reduce Medicare waste without a patient-safety crisis. The accountability design would become the reference for every federal health programme that follows.

Without the liability framework, WISeR accumulates its risk quietly. Not through a single dramatic incident, but through the gap between AI recommendation volume and human review capacity, compounded by an accountability vacuum no published document has yet closed. That gap does not stay open indefinitely.

Pre-Mortem: The US Government’s 3,611 AI Use Cases

Posted on by

On 3 April 2025, the White House issued OMB Memorandum M-25-21, directing every major federal agency to appoint a Chief AI Officer, expand the use of artificial intelligence across government operations, and manage risk proportional to each system’s impact on citizens. Twelve months later, the Federal Agency AI Use Case Inventory records 3,611 AI use cases across 56 agencies, more than double the prior year’s total. A May 2026 survey of more than 200 technology executives across civilian and defence agencies found 53% are actively planning agentic AI pilots. Only 8% of those agencies have incident response frameworks in place.

This is the eleventh piece in the Pre-Mortem series. Five questions, applied to the public record, before a programme has had the chance to succeed or fail.

 

The Bet

The US government is betting that embedding AI across 3,611 federal workflows covering benefits decisions, immigration adjudications, healthcare determinations, and law enforcement, will make government faster and more efficient before the accountability architecture governing those decisions is clarified. OMB M-25-21 requires Chief AI Officers, public AI inventories, and risk management proportional to impact. The hard compliance deadline for role-specific AI training arrives in September 2026. If that architecture catches up to the deployment before a consequential wrong decision reaches a citizen with no named relief, the bet holds.

 

The Assumption

The expansion’s credibility turns on one unanswered question: whether the Federal Tort Claims Act, designed to govern negligent acts by human federal employees, applies without amendment to decisions made by AI agents running inside federal systems. The same May 2026 survey found only 44% of agencies include vendor liability clauses in AI contracts, and only 29% have documented kill-switch procedures. The legal architecture governing accountability in federal government was designed for humans acting on behalf of the state. No court has ruled on whether it extends to the agents they built.

 

The Sequence

In 2024, federal agencies reported 1,757 AI use cases. By 2025, that figure had grown to 3,611. In March 2026, the Department of Veterans Affairs expanded AI use in claims processing, with 215 of its 367 AI systems classified as high-impact, covering benefit eligibility, healthcare access, and fraud detection. In May 2026, the majority of agencies were planning agentic pilots, with only 20% having defined pre-deployment testing policies. The AI reached citizens before the accountability reached the AI.

 

The Pager

Russell Vought, Director of the Office of Management and Budget, carries the M-25-21 mandate at the centre of the federal AI expansion. Every covered agency has designated a Chief AI Officer responsible for inventory, risk management, and AI governance at agency level. The VA alone runs 215 high-impact AI systems. No single published document names what relief is available to a veteran whose claim was influenced by one of those systems, which official carries accountability for that decision, or whether the Federal Tort Claims Act applies when the acting party is software, not a civil servant.

 

The Proof

The measure that would settle this is a published legal standard: a named accountability chain clarifying who carries liability when a federal AI agent makes a consequential wrong decision, whether government, vendor, or joint, and whether the Federal Tort Claims Act applies or new legislation is required. No such standard has been published. OMB M-25-21 requires risk management proportional to impact. It does not name the relief available to a citizen when that risk management fails, nor the date by which that question must be answered.

 

Verdict

If OMB publishes, before the September 2026 training compliance deadline, a named accountability standard for AI-driven decisions in high-impact federal systems, covering who carries liability when the AI is wrong and what legal remedy a citizen holds, the expansion will stand as the most deliberate attempt the US federal government has made to govern AI before it reaches citizens at scale. Without that, the US government has put AI into 3,611 workflows and left the question of who carries the call when the AI gets it wrong to be answered in court, by accident, or not at all.

Regional Leadership Roles Reward a Different Kind of Patience Than Western Corporate Careers Teach.

Posted on by

Most Western executives arrive in a regional leadership role with the instincts that got them promoted at home: move fast, show visible wins early, let the quarterly numbers do the talking. Those instincts do not transfer. They actively work against you.

Western corporate careers train impatience, and reward it. Promotion cycles run on twelve to eighteen month windows. Performance reviews measure the quarter just closed. The fastest route up is a visible individual win, attached cleanly to your name, delivered before the next review cycle starts. None of that is wrong inside the system that built it. It is a system optimised for institutional and contractual trust, built on the assumption that the relationship survives even if the individual leaves tomorrow.

 

What Regional Leadership Rewards

Gulf leadership runs on a different clock, and the data on how regional leaders actually spend their time backs that up. PwC’s most recent Middle East CEO survey found Saudi Arabian chief executives dedicate 27% of their time to planning five years or more ahead, compared with 16% globally, and GCC chief executives overall spend meaningfully more of their schedule on long-term strategic planning than their global counterparts. That time allocation is a rational response to a business environment where relationships carry the weight of a deal through its inevitable rough patches.

The mechanism underneath that patience has a name: wasta, the use of personal relationships and trust networks to get things done. Research summarised by Northeastern University’s D’Amore-McKim School of Business puts it plainly: “in the Arab context, trust at the interpersonal level needs to be established before any business relationship can unfold.” That is not a bureaucratic delay. It is the actual mechanism through which decisions get made, and skipping it does not speed anything up. It just means the decision never fully lands.

 

The Failure Rate Everyone Quotes Does Not Hold Up

Ask any expat executive circle and you will hear it: 20 to 40% of international assignments fail. It is one of the most repeated statistics in expatriate management, and it is very likely wrong. Anne-Wil Harzing’s 1995 analysis traced the figure back to a small number of frequently misquoted articles, only one of which actually contained solid empirical evidence, and that evidence showed failure rates to be low. The myth persisted for three decades anyway, because it made a better opening line for a consulting pitch than the more boring truth.

The more useful question is why the ones who do fail actually fail. The qualitative research on that is far less contested: misjudging the local pace of trust-building rather than any shortage of competence or effort.

 

What Patience Looks Like in Practice

Cross-cultural consultancy Commisceo Global puts the practical advice plainly: do not expect deals to be completed in two visits to the region, because the relationship has to be nurtured before the commercial conversation can move forward at all. That advice sounds soft to an executive trained to treat a second meeting without a signed term sheet as a stalled deal. It is a description of how decisions actually get authorised in a hierarchy where personal trust in the room outweighs the org chart on paper.

Twenty years running programmes across markets that operate on this clock, the pattern I have seen most often has nothing to do with understanding the concept. Most executives can recite the advice back accurately in an interview. Where they struggle is stopping themselves from behaving as though the Western clock is still running underneath it: checking for the win they would expect by month three, reading the absence of one as a signal that the relationship has stalled, when it has not.

 

Three Things That Actually Help

Building trust deliberately, rather than waiting for it to accumulate on its own, takes specific behaviour. Spend real time in the room before asking for the decision. Learn who actually holds influence in a given relationship, because it is rarely only the most senior title at the table. And resist manufacturing a visible early win for the story it tells back home, because the fastest way to damage the trust being built is to be seen optimising for your own next review cycle instead of the relationship in front of you.

 

What This Requires

The patience regional leadership rewards has nothing passive about it. Ninety-three per cent of CEOs across the GCC expect domestic momentum to keep strengthening, against 55% of their global peers. These are not leaders coasting on a slow market. The patience is aimed at a specific target: building the trust that makes fast execution possible later, instead of chasing a fast, visible win the relationship cannot yet support.

That is a genuinely different skill to the one a Western corporate career trains for, and it rarely shows up on the resume that got someone the role in the first place. The executives who actually succeed in regional leadership rarely arrive already patient. They learn to notice, quickly, that the clock they were trained on is the wrong clock, and recalibrate before the org chart tells them they have run out of time to.