There are two ways to position a risk function inside a financial institution, and the choice between them determines almost everything else about how it works.
The first is the gatekeeper. Risk sits at the end of a process, reviews what the business has already decided, and either approves or declines. Its value is measured by what it stops. The second is the architect. Risk sits at the beginning, shapes the design of products and decisions while changing them is still cheap, and builds governance structures the institution actually uses. Its value is measured by what it makes possible.
Most risk functions live somewhere between these two positions. Very few choose where. The position is usually inherited from structure and incentive, not decided on purpose. That is the problem worth examining.
Why the structure pulls toward gatekeeper
Left to its own incentives, a risk function drifts toward the gatekeeper end. The drift is not a failure of ambition. It is a response to what the institution can see and credit.
Saying no is visible. A declined proposal, a blocked campaign, a condition attached to an approval: these are legible events. They appear in minutes. They can be pointed to later as evidence that the function was doing its job. Shaping a product upstream is the opposite. When risk influences a design early, the result is a better decision that never becomes a visible event. There is no record of the bad version that was avoided, because it was never built. The contribution is real and almost impossible to credit.
An institution rewards what it can measure. It can measure refusals. It struggles to measure the quiet improvement of a decision that happened before the decision was formally made. So the incentives push the function toward the activity that shows up, and the activity that shows up is gatekeeping.
The cycle this creates
Once a risk function is experienced primarily as the place where things get stopped, a self-reinforcing pattern sets in, and it works against everyone in it.
The business learns that bringing something to risk early invites questions rather than help. So it brings things late, when the design is largely settled and the cost of changing it is high. Risk, handed a near-final proposal, can no longer shape it. The only tools left are conditions and refusals. So it attaches conditions. The business experiences this as friction that confirms its original instinct. So it resolves to involve risk even later next time. Each turn of the cycle moves risk further from the design and closer to the gate.
The outcome is a function that is structurally prevented from doing its most valuable work, while being blamed for the friction that the structure itself produced. Nobody designed this. It assembles itself out of rational responses to bad incentives, which is exactly why it is so durable. You cannot fix it by asking people to behave differently inside it. The incentives will pull them back.
What the architect position actually requires
It is tempting to treat architect versus gatekeeper as a matter of attitude, as though a risk function could simply decide to be more collaborative. It cannot. The difference is structural, and it lives in one specific thing: timing.
Institutions where risk genuinely functions as an architect share a single characteristic. The risk team is involved early enough that its input changes the design, not just the approval timeline. That is the whole distinction. Early enough to shape what gets built is architect. Late enough to only judge what was built is gatekeeper. Everything else follows from where in the process the function sits.
Moving risk earlier is harder than it sounds, because it requires the business to give something up. Early involvement means letting risk into the room before the design is settled, when ideas are still soft and ownership is still forming. That feels like surrendering control of the creative part of the work to the function most associated with constraint. The instinct to resist it is understandable. It is also the precise instinct that keeps the cycle running.
It also requires risk to change what it offers in that room. A function that arrives early and behaves like a gatekeeper, leading with objections to a design that does not yet exist, teaches the business that early involvement is worse than late involvement. The architect position has to be earned by being useful upstream: helping shape a workable version rather than only naming the problems with an unworkable one. That is a different skill from review, and not every risk function has built it.
The choice underneath the choice
Architect or gatekeeper is usually framed as a question about the risk function. It is better understood as a question about the institution.
A function does not become an architect on its own. It becomes one when the institution decides that risk input is worth having early, builds the relationship that lets it in, and changes what it rewards so that shaping a decision counts as much as stopping one. None of that is within the gift of the risk team alone. It is a choice the institution makes about what it wants its risk function to be, and then backs with structure.
The institutions that make this choice well do not end up with a softer risk function. They end up with a more demanding one, positioned where its judgment can actually change outcomes rather than just register them. The approval at the end means more, because the design that reached it was shaped by the same discipline that will judge it. The institutions that never make the choice are left with a capable function held at the gate, measured by its refusals, and quietly prevented from doing the work that would have been worth the most.
What determines whether a risk function is shaping decisions or reacting to them is not the talent inside it. It is where the institution has decided to let it stand.
I write about governance, risk, and the decisions institutions find hardest to make. If this is relevant to a problem you are working through, reach me at aan@asifahmednoor.com.
← Back to writing