Morgan Stanley has taken a novel approach to automating one of its most critical financial tasks: profit and loss reconciliation. By reducing the autonomy of its AI agents, the firm has managed to halve the time spent on this complex task. This unexpected strategy highlights the bank’s focus on maintaining human oversight and accountability in AI-driven processes, a move that sets it apart in an industry often chasing full automation.
## FIXR Behind the Scenes
Morgan Stanley’s FIXR system is at the heart of this transformation. Each trading day, the firm’s trade desks engage in transactions involving cash equities and debt investments. At day’s end, controllers face the daunting task of reconciling profit and loss across various systems, a process fraught with mismatches in data. Previously, controllers could spend up to six hours manually resolving these discrepancies for a single book. With FIXR, this task now takes only two to three hours, saving around 1,500 hours weekly across 100 controllers.
FIXR operates by analyzing post-calculation “breaks” and suggesting resolutions based on established rules. It employs multiple agents to learn from past resolutions, codifying these into automated rules over time. While the system can auto-clear familiar breaks and suggest solutions, it defers to human controllers for unfamiliar situations, ensuring every decision is vetted and fed back into the system for continuous improvement. According to Morgan Stanley Managing Director Todd Johnson, this iterative learning process is crucial for building trust and ensuring accountability.
## Competitive Context
In the broader financial sector, AI deployment often aims for full autonomy, with firms racing to automate as many processes as possible. Morgan Stanley’s approach diverges from this norm by emphasizing a human-in-the-loop model. While other institutions might view human oversight as a bottleneck, Morgan Stanley sees it as a necessary component for effective and reliable automation.
This strategy could serve as a model for other enterprises grappling with the challenge of integrating AI into their operations without sacrificing accuracy and accountability. By focusing on process intelligence and mapping workflows before implementing AI, Morgan Stanley ensures that automation is applied where it is most effective, rather than forcing technology into unsuitable roles.
## Real Implications for Founders, Engineers, and the Industry
For founders and engineers, Morgan Stanley’s experience with FIXR underscores the importance of balancing automation with human oversight. The key takeaway is that AI should augment human capabilities, not replace them entirely. This approach not only preserves jobs but also enhances the quality of the work produced, as human expertise remains a critical component of the decision-making process.
For the industry at large, Morgan Stanley’s methodology offers a blueprint for integrating AI in a manner that prioritizes trust and accountability. As AI continues to evolve, the financial sector may see a shift towards similar models that leverage human expertise to guide and refine automated systems. This could lead to more robust and reliable AI applications, as well as a more gradual and sustainable integration of technology into sensitive workflows.
## What Happens Next
Morgan Stanley’s success with FIXR suggests that the future of AI in finance may lie in hybrid systems that combine human oversight with machine efficiency. For founders and engineers, this means considering how AI can complement human skills in their own projects. The challenge will be to design systems that are not only efficient but also transparent and accountable, ensuring that human judgment remains a core component of complex decision-making processes.
