Robin Hasson, Head of Reconciliations – Product Management at Smartstream, discusses a major evolutionary leap in financial services technology: the industry’s shift from simple process automation to full autonomy whereas automation typically focuses on point-to-point processes, such as loading data or managing throughput.
Autonomy, conversely, is defined by chaining multiple automated processes together into much longer, goal-driven workflows. This fundamentally changes the system’s purpose, enabling it to autonomously derive and execute the single best way to solve a complex problem rather than just following a pre-set routine. This new approach represents a significant evolution from the tools provided over the last decade, which simply allowed people to manually find, report, and deal with data themselves. Now, Smartstream is focused on an “outcome first” model where the system proactively seeks out problems before the user even knows they exist. Beyond just flagging an issue, the autonomous system is designed to propose an out-of-the-box solution, thereby delivering the end goal directly to the user and eliminating the tedious manual search for data or workarounds.
The deployment of these highly autonomous systems, however, requires careful design, particularly in the heavily regulated environment of financial services. Smartstream cautions that firms cannot simply allow a “black box” to control everything; systems must be designed to meet strict regulatory and business process needs. Accountability is paramount: there must be full traceability, audit, and control, including data lineage that tracks every change and explains the logic behind every decision. Crucially, the human element remains accountable and humans must retain the ability to sign off, approve, reject, or control the system’s actions, ensuring they can prove to regulators exactly what transpired at any point in the future
Looking at the world of agentic workflows, Hasson acknowledged that while agents can handle tasks like designing workflows, orchestrating connections, and simulating user activities, the challenging part is accurately capturing real business intent. This deep domain knowledge is something financial firms and providers like Smartstream have spent years fine-tuning with clients. Smartstream’s strategy is to leverage this existing, highly accurate knowledge, rather than starting from scratch, to build agentic workflows, resulting in greater value and more trustworthy systems. Smartstream anticipates that while this technology is in its early stages, it could eventually automate most back-office functions, provided the proper levels of human scrutiny, control, and accountability are in place.
