Automate Thyself

The Coming Cognitive Shift in Financial Services

Around 2016, BlackRock introduced Aladdin Risk for Wealth Management, an adaptation of its institutional analytics engine tailored to financial advisors. The idea was simple enough: democratize access to institutional-grade portfolio construction tools. Over a decade later, Aladdin still sits at the center of BlackRock’s technology narrative, and for good reason. It captures the logic of an industry in transition - from people making decisions, to systems recommending them.

But if Aladdin was the beginning, we’re now approaching something of an endgame or the middle...

From Succession

The debate about automation in financial services has largely been confined to two areas: cost and compliance. The former drives operational efficiency; the latter, reputational risk. Both are important, but both are incremental. The next stage, however, is about something far more fundamental: replacing cognition itself.

This may sound like a lot. Trust me, it is not.

A Brief History of Judgment

Financial services, particularly asset and wealth management, have always styled themselves as businesses of judgment. The job of an asset manager, at least historically, was to synthesize information, develop theses, allocate capital, and communicate rationale. All of this happens within a narrow regulatory envelope, but it is, at core, a cognitive function.

There’s a reason investment firms have long drawn from the same intellectual gene pool as consulting and law. They are institutions of thought. But they are also institutions of process. And this is the critical distinction that artificial intelligence has begun to exploit.

Large language models are not infallible. But neither are humans. The difference is that LLMs are cheap, scalable, and  crucially - tractable in code. Human intuition is difficult to model and impossible to batch process. As AI becomes better at mimicking high-level cognitive tasks, the locus of value in finance will begin to shift - away from bespoke human thinking and toward scalable, self-improving systems of reasoning.

---

---

The Cost of Incumbency

It is tempting to think that large incumbent firms are best positioned to benefit from AI. After all, they have the data, the distribution, and the balance sheet to deploy meaningful capital. But incumbency in finance is often as much a liability as an asset. Systems are complex, workflows are fragile, and the cultural cost of change is high.

There is a reason core banking platforms have proven difficult to replace, even when superior technology exists. The same can be said of trading systems, CRM stacks, risk engines, and onboarding flows. Each has grown around layers of legal constraint and human habit, intertwined in ways that defy easy refactoring.

To leverage AI in a truly transformative way, institutions must be willing to rethink not just their tech stacks, but the assumptions embedded in their organizational structure: Who gets to decide? How are decisions surfaced and recorded? What happens when compliance oversight shifts from retrospective review to embedded real-time constraint?

None of this is impossible, but it is difficult. Which is why the real advances are likely to come from the edges.

Intelligence as a Platform

In financial services, the dominant platforms of the last twenty years have been built on infrastructure and access: think Rupay, Razorpay, account aggregators. They sit between the user and the rails. But as generative AI matures, we will see the rise of a new class of platform: one that mediates judgment.

Consider a future where every client relationship is managed by an LLM-based assistant that knows the client’s entire history, preferences, tax situation, investment philosophy, and communication style - and updates these in real time. This is already being prototyped.

The implications are profound. If the advisor is software, the firm becomes a systems integrator. The value shifts from branding and personalities to design and control of the AI interface itself. It is less about what you believe the client should do, and more about what logic trees and datasets you embed in the assistant’s worldview.

In such a world, financial advice becomes less artisanal and more like search: a commoditized front-end to a highly complex back-end. Brand differentiation will come not from alpha, but from architecture.

---

---

The Regulatory Imagination

None of this will occur in a vacuum. Finance is heavily regulated and rightly so. But regulators, too, face a choice.

They can continue to supervise the human processes they understand - the investment memos, the KYC forms, the model portfolios. Or they can begin to develop frameworks for supervising machine reasoning - which is probabilistic, iterative, and non-deterministic.

This is not without precedent. Risk-based capital models already accept a level of statistical abstraction in how banks represent asset quality. But AI introduces a level of opacity and adaptation that is hard to reconcile with existing regimes of accountability. What does fiduciary duty mean when a synthetic persona is making the recommendation?

From Succession

In time, regulators will have to answer these questions. And firms that can anticipate the likely answers by embedding auditability, explainability, and constraint into their AI systems will enjoy a first-mover advantage.

Automation as Strategy

The real strategic question is this: how much of your firm’s value is embedded in processes that require human cognition, and how much of that cognition can be replicated at scale?

If the answer is “a lot” — and the cost of replicating it continues to fall — then the obvious conclusion is that much of your firm is ripe for automation.

This creates a paradox. To preserve value, you must automate. But to automate, you must destroy the very structure that once captured the value.

It is here that strategy becomes self-reflexive. The firms that survive the transition will be those that are willing to automate themselves - not as a cost-cutting measure, but as a philosophical commitment. They will view every human process as suspect, every cognitive workflow as a candidate for refactoring, every piece of intellectual capital as a training set in waiting.

From Succession

This is not easy. It demands conviction, humility, and vision. But the alternative - defending cognitive territory while the ground shifts beneath your feet is harder still.

Epilogue

The first revolution in finance was mechanical: we digitized paper. The second was transactional: we disintermediated access. The third was informational: we built platforms to synthesize data.

We are now entering the fourth: the cognitive revolution. The business of finance at its core, is the business of reasoning under uncertainty. And that, increasingly, is a job for machines.

---