Written by: Ashraf Mohamed, Manager, AI Service Engineering
The bonds are genuinely weakening
It has become common to reassure the financial industry that AI will not change the fundamentals of the customer relationship. That reassurance deserves skepticism. The bond between financial institutions and their customers, built over decades of branches, relationship managers, and paperwork, is genuinely becoming brittle, and agentic AI is a direct cause. This brittleness is not a side effect to be patched; it is built into how the technology works. It is also, as we will see, only half of the story. An autonomous agent that can compare rates, negotiate fees, execute KYC checks, and move money on a customer’s behalf does not merely automate a task; it inserts an independent actor between the bank and the person it serves.
The World Economic Forum’s 2026 research across more than 150 senior financial executives found that as agentic systems take on more autonomous decisions, questions of accountability and customer trust are intensifying faster than anticipated. McKinsey’s 2026 AI Trust survey describes the same shift: the industry has moved from worrying about AI saying the wrong thing to worrying about AI doing the wrong thing, autonomously and at scale. In physical terms, the system is losing binding energy, and pretending otherwise is neither good analysis nor good strategy.
But systems that lose a force rarely just disperse
Here is where the prevailing narrative stops too early. In physics, the loss of a dominant binding force does not necessarily cause a system to fly apart. When other interactions remain significant, the system can undergo a transition and settle into a new equilibrium governed by the interactions that become dominant under the new conditions. The disintermediation of the old customer relationship is real, but it is not the whole picture. The strategic question for financial institutions is not “how do we stop the old bonds from weakening?” This shift is unlikely to be reversed. The real question is “what is the new dominant force, and who will control it?”
The core idea: the combination is the new center of gravity
The answer is not agentic AI alone. There is another, transformative force coming out of the dark sooner or later, one that will unlock new frontiers and make life even more autonomous. Yes, you got it right: quantum computing. The coupling between these two new forces is the core idea that deserves more attention than either technology receives on its own.
Agentic AI creates the demand. Autonomous agents acting at machine speed still need somewhere to settle, verify, and be held accountable. Recent International Monetary Fund (IMF) analysis on agentic payments makes this explicit: as independent agents initiate financial actions, the trust layer (identity verification, provenance, and policy-based access control) becomes more valuable [1]. Whoever operates that layer becomes the anchor that lets probabilistic, autonomous systems act inside a framework with legal certainty.
Quantum computing has the potential to supply the computational capacity that this new role may eventually require. An agentic financial system generates decisions at a rate and complexity the old back office never faced: real-time fraud scoring across millions of agent-initiated transactions, portfolio optimization under dozens of constraints, risk simulations that reprice by the second. Quantum computing is among the most promising computational paradigms for tackling certain classes of optimization and simulation problems as they continue to scale. According to “The Quantum Insider”, more than fifteen global banks already run active quantum research programs targeting these problems, and early results are no longer hypothetical: a 2026 backtest using quantum-inspired optimization on a ten-equity portfolio produced a Sharpe ratio of 1.81 against 1.31 for simulated annealing [2].
Neither force does the work alone. Orchestration without computational depth cannot keep up with the decision rate it is supposed to govern; computational depth without the trust layer has nothing to anchor to. Together they behave like a coupled system whose combined field defines where the industry’s mass will settle. That coupling is the new center of gravity. Institutions gravitate toward it.
What this mean on a leader's desk
For decision-makers, this reading translates into three commitments:
- Sequence capability, not technology. The question isn’t “when do we buy quantum?” but “which of our problems are combinatorial optimization in disguise?” Reconciliation, collateral allocation, and liquidity routing all qualify. Formulate them now, and switching backends later becomes easy.
- Make governance a promotion path. Treat agent oversight as a compliance chore assigned to whoever is free, and the best engineers will avoid it, leaving the institution’s most valuable future capability staffed by accident.
- Protect a small, real quantum literacy effort. Not a lab, not a press release: a handful of engineers with real time to build intuition on actual business problems. The edge won’t go to who spent the most, but who can tell signal from noise when procurement decisions arrive.
What we are doing at Finaira
This is not an abstract thesis for us. It is the premise behind how Finaira thinks about the years ahead, and it shapes where we choose to build depth rather than breadth. We are deliberate about which capabilities get invested in early, precisely because the institutions that treat this coupling as a single, long-term commitment rather than two separate bets will be the ones still standing at the center when the field settles.
References
- IMF – “How Agentic AI Will Reshape Payments,” IMF Notes 2026/004: https://www.elibrary.imf.org/view/journals/068/2026/004/article-A001-en.xml
- Mancilla et al. Constrained Portfolio Optimization via Quantum Approximate Optimization Algorithm (QAOA) with XY-Mixers and Trotterized Initialization: A Hybrid Approach for Direct Indexing. arXiv:2602.14827 (2026)