Issue #001 introduced Layer 1 — Purpose. The mandate an institution is founded on. The reason it exists. Non-negotiable, until it quietly stops being treated that way.
Issue #002 examined Layer 2 — Strategy. How Purpose meets competitive pressure. How a board can approve a strategy that technically succeeds — and structurally destroys the institution it was meant to serve.
This issue moves to Layer 3. Where a directive leaves the board and enters the organisation. And where something gets lost in translation.
The institutions referenced in this issue are cited on the basis of publicly documented regulatory findings, official investigations, and other publishd reports. All analysis is educational. Nothing here constitutes legal, regulatory, financial, or investment advice.
Purpose tells an institution why it exists. Strategy tells it where to compete. Intent is the translation layer — the point at which a board mandate becomes an institutional instruction.
When Intent is precisely defined, execution follows the mandate. When Intent is loose — when a mandate is issued without specifying its boundaries, its constraints, what it permits and what it prevents — the institution fills the gap. Not with malice. With interpretation.
In both cases in this issue, the mandate was legitimate. The goal was sound. Nobody was instructed to cause the outcome that followed.
The Intent was simply never engineered with enough precision to prevent it.
The board said one thing. The institution heard another. Both believed they were aligned. Neither checked whether the space between the mandate and the execution had been closed.
This issue examines what fills that space — and what defining Intent precisely would have changed.
Wells Fargo is one of America's largest retail banks, serving tens of millions of customers across the United States. Its Purpose is straightforward and long-standing: serve customers well, build lasting relationships, and grow through genuine engagement rather than acquisition. Conservative in lending, resilient through the 2008 crisis, it built its reputation on the idea that a bank that truly served its customers would naturally deepen those relationships over time.
Its Strategy reflected that Purpose directly. Cross-selling — offering existing customers additional products they would benefit from — was the natural growth lever. Deeper relationships meant more products per household. The number that crystallised this was eight. Eight products per customer. "Eight is great" became both a strategic slogan and a performance target across the retail network.
Out of that Strategy came a mandate. Senior leadership set an Intent to operationalise cross-selling across every branch, every channel, every customer interaction. The mandate was genuine: deepen relationships. Serve customers more completely. The Intent behind it was never in question.
What the mandate never defined was its boundary. Where does encouraging a customer to open a product they would benefit from end, and opening a product on their behalf — without their knowledge — begin? The mandate said what the institution wanted. It never said what it would not permit in pursuit of it.
That question was left to the institution to answer. Under branch-level sales targets, performance rankings, and incentive structures tied directly to the eight-product metric, the institution filled the silence with its own logic.
Between 2011 and 2016, Wells Fargo employees opened approximately 3.5 million unauthorised accounts. Credit cards customers never requested. Insurance products attached without consent. Fees charged on accounts that existed only because a target needed filling. 5,300 employees were terminated for conduct that was, in every case, consistent with a rational interpretation of an ambiguous mandate.
No one at board level instructed fraud. The cross-selling mandate was real. But the mandate was issued without defining where it ended. Execution filled that space — and the space was wide enough for 3.5 million decisions to accumulate inside it, none individually sanctioned, all of them collectively catastrophic. Wells Fargo paid $185 million in regulatory fines in 2016, followed by a $3 billion settlement with the DOJ and SEC in 2020. The CEO resigned. The reputational damage took years to quantify and has never been fully repaired.
Kaiser Permanente is one of the largest integrated health systems in the United States, serving over 12 million members across eight states. Its Purpose is foundational to its structure: deliver comprehensive, high-quality care by combining insurance and care delivery under one roof. Prevention over treatment. Continuity over episodes. The long-term health of the patient as the governing principle — not the episodic transaction.
Its Strategy reflected that Purpose directly. Integration — across specialties, across care settings, across the patient's lifetime — is the competitive and clinical logic of the Kaiser model. Mental health, historically siloed from physical care across the US system, was an obvious gap. Treating it as a separate, stigmatised referral pathway was inconsistent with what Kaiser existed to provide.
Out of that Strategy came a mandate. Leadership set an Intent to integrate mental health care into primary care — to make behavioural health as accessible as any other medical need, remove the friction of a separate referral pathway, and meet patients where they are. The mandate was genuine. The aspiration was entirely consistent with Kaiser's Purpose.
What the mandate never defined was what integration required at the operational level. Integrate how? To what standard? When a primary care physician identifies a mental health need, what happens next — and within what timeframe? What is the minimum acceptable waiting time for a follow-up specialist appointment? At what point has the mandate been violated rather than served?
Those questions were left unanswered. The institution — managing costs, appointment volumes, and clinical capacity — interpreted the mandate in the language of what it could control. Integration became a reduction in specialist referrals. Meeting patients where they are became routing complex mental health needs through primary care settings that were not equipped to handle them.
A 2023 investigation by the California Department of Managed Health Care documented the result: inadequate appointment availability, patients waiting weeks or months for follow-up care, and a structure that had made mental health services harder to access than before the mandate was issued. Approximately 50,000 patients were identified as having been inadequately served. Kaiser paid a $50 million fine — the largest penalty in the department's history at the time. A subsequent settlement required the hiring of hundreds of additional mental health clinicians, significant new infrastructure investment, and ongoing regulatory monitoring.
The mandate was not abandoned. It was genuinely believed throughout. But it was never made precise. The institution filled the silence with cost and capacity logic — rationally, given what it had been told to optimise for. And 50,000 patients experienced the distance between an aspiration and a defined standard.
Different sectors. Different scales. Identical structural failure.
In both cases, a legitimate mandate was issued by leadership. It was genuinely meant. Nobody was instructed to produce the outcome that followed. And yet the outcome followed — consistently, at scale, over years.
The reason is not misconduct. The reason is that the mandate was issued without boundaries. In both cases, leadership was precise about what the institution wanted — deeper customer relationships, integrated mental health care — and entirely silent about what the mandate protected, what it prevented, and what would constitute its violation.
When a mandate is silent on its own constraints, the institution fills the silence. Not randomly. It fills the silence with the logic of whatever pressure is most present. At Wells Fargo, that pressure was the sales incentive structure tied to the eight-product target. At Kaiser, it was cost and capacity management. In both cases, the filling was rational — given what the institution had been told to optimise for. The problem was not that people made bad decisions. The problem was that the mandate gave them no basis for making the right ones.
Intent without boundaries is not a directive. It is an invitation for the institution to decide what the mandate means.
In the Decision Integrity Chain™, the Layer 3 failure opens here. The board sets a mandate. The institution executes against it. But the space between the mandate and the execution is ungoverned — no precision on constraints, no operational definition, no threshold below which the mandate has been violated rather than served. That space is where 3.5 million unauthorised accounts accumulate. Where 50,000 patients wait months for care they were promised.
The Layer 3 failure is not visible at the moment the mandate is issued. It sounds right. The aspiration is genuine. The failure only becomes visible when execution accumulates — when thousands of individual decisions, each consistent with a reasonable interpretation of an ambiguous mandate, produce an aggregate outcome that no board would have sanctioned as a single choice.
That is what makes Intent the most dangerous layer. Purpose can be tested against the founding mandate. Strategy can be reviewed against the competitive environment. Intent is invisible until it has already done its damage — because by the time the outcome is visible, the mandate that produced it has long been treated as settled.
The fix for a Layer 3 failure is not better monitoring after the fact. It is Intent engineering before execution begins. A mandate that cannot answer three questions is not ready to be issued: What does this mandate permit? What does it prevent? And at what point has the institution's execution of it failed the mandate rather than served it?
Before the cross-selling mandate was operationalised into branch targets, one question should have been formally answered: When does offering a customer an additional product cross the line into opening a product they did not ask for?
In practice, answering that question would have produced one rule: no account may be opened without the customer's explicit, documented initiation. Not implicit consent. Not inferred interest. A recorded request or a signed application — nothing else. That one boundary, written into the mandate at the point of issue, closes the space in which 3.5 million decisions were made.
The cross-sell target could have remained. Eight products per household is an ambitious but legitimate goal if each product reflects a genuine customer need. What the boundary removes is the ability to reach the target by manufacturing accounts rather than earning them. The metric stays. The abuse of the metric becomes structurally impossible.
Branch managers operating under that boundary would have faced a different pressure: not how many accounts can we open today, but how many customers can we genuinely serve today. The incentive structure would have required redesign to match. That redesign is a governance conversation. It costs a week of leadership time. The absence of it cost $3.185 billion in fines, the resignation of a CEO, and a reputational recovery that is still ongoing.
Before the mental health integration mandate was deployed operationally, one question should have been formally answered: What does a patient actually experience if this mandate is working — and what do they experience if it is not?
In practice, answering that question would have produced a floor: no patient flagged by a primary care physician as requiring mental health follow-up waits more than a defined number of days for that appointment. A specific number — not a target, not an aspiration, but a minimum standard below which the mandate has been violated. That floor, written into the mandate before it reached operational planning, changes what cost and capacity management is permitted to do. Managers can find efficiencies above the floor. They cannot find efficiencies by lowering it.
The integration goal could have remained exactly as stated. Mental health in primary care is the right clinical direction. What the floor prevents is the interpretation of integration that reduces rather than improves access — the version that routes complex needs through settings unequipped to handle them, because that is cheaper and faster to manage. The mandate's aspiration stays intact. The execution path that violated it becomes non-compliant by definition.
A clinical governance conversation at the point of mandate issue — establishing the floor, naming the access standard, defining what violation looks like — would have taken days. The investigation, the $50 million fine, the regulatory monitoring, the infrastructure remediation, and the impact on 50,000 patients took six years.
Every mandate has a direction — what the institution is trying to achieve. Intent engineering requires a second document issued at the same time: the boundary definition. What does this mandate permit? What does it explicitly prevent? What actions are incompatible with it even if they advance the stated metric? A mandate that cannot name what it will not permit has not been engineered. It has been announced.
At Wells Fargo, the Intent Boundary Statement would have contained one line: no account may be opened without explicit, documented customer initiation. That single sentence closes the space in which 3.5 million decisions were made. The cross-sell target remains. The mechanism for abusing it does not.
At Kaiser, the Intent Boundary Statement would have contained one line: integration does not mean reduction in specialist access — any routing through primary care that results in a patient waiting longer than they would have under a direct referral constitutes a failure of this mandate, not an execution of it. That one sentence makes the misinterpretation non-compliant by definition before it begins.
The target tells the institution what success looks like. The floor tells it what failure looks like — the threshold below which execution has violated the mandate rather than served it. The floor is set at the same moment as the target, by the same authority that issued the mandate. No target is issued without its floor. Targets without floors are pressure without limits. The outcome is what pressure finds when there is nothing to stop it.
At Wells Fargo, the floor would have been a customer consent standard: every product counted toward the eight-product target must be traceable to a documented customer interaction in which the product was discussed and accepted. Any account not meeting that standard is below the floor — regardless of the number it contributes to branch performance.
At Kaiser, the floor would have been a waiting time standard: no patient flagged for mental health follow-up waits more than a defined number of days for that appointment. That number — whatever the clinical governance team determined was clinically appropriate — becomes the non-negotiable lower bound. Cost management operates above it. It does not operate instead of it.
Before any significant mandate is operationalised, it should pass one structured question: given a realistic set of pressures — cost constraints, sales targets, capacity limits — what is the most rational thing a manager could do while technically complying with this mandate? If the answer includes outcomes the board would not sanction, the mandate is not ready. The Intent has not been engineered.
At Wells Fargo, the pressure test answer would have been immediate: under branch-level performance rankings tied to eight products per household, the most rational thing a manager can do is open accounts against any customer record with an open product slot — regardless of whether the customer asked. That answer, surfaced in a governance conversation before deployment, changes the mandate. It does not change after 3.5 million accounts have been opened.
At Kaiser, the pressure test answer would have been equally clear: under cost and capacity pressure, the most rational thing a clinical operations manager can do is route mental health needs through primary care — because it is cheaper, faster, and technically compliant with a mandate that says "integrate into primary care" without defining what integration requires. That answer, surfaced before deployment, adds the floor. It does not add it after 50,000 patients have experienced the gap.
None of these require new technology. None require new regulation. All three require a board that treats the precision of its mandates as a governance responsibility — not a communication preference.
That is Decision Engineering™ at Layer 3.
When your institution last issued a significant mandate — a target, a strategic priority, an operational change — did anyone document what that mandate will not permit? And did anyone run the pressure test: given the targets attached to this mandate, what is the most rational thing a manager can do while technically complying with it?
If those questions were not asked, your institution has already decided what the mandate means. It just has not told you what it decided yet.