AI Answer: Solving Reconciliation Challenges

Practical artificial intelligence (AI) use cases and applications in buy-side reconciliation and post-trade operations are prevalent, yet many have yielded uneven and questionable results. AI can go only so far to autonomously solve the operational problems investment management firms face. The solution? An intelligence augmentation (IA) approach to AI.

Artificial intelligence (AI) is a ubiquitous term that can mean different things to different people. At its basic level, AI mimics human cognition by independently processing, reasoning and creating results faster than a person can. While some progress has been made in the financial industry, one cannot assume that AI and machine learning (ML) algorithms will fix every problem – especially complex problems investment managers face every day.

Instead, people must always play a key role in the AI equation to create intelligence augmentation (IA) to yield the most benefit. Although the underlying technologies that fuel AI and IA appear identical, their approaches to meeting goals can be very different. AI looks to build solutions without human involvement, while IA attempts to build solutions that make humans better at their job.

Why Intelligence Augmentation?

Also known as intelligence amplification, cognitive augmentation, machine augmented intelligence and enhanced intelligence, IA has a long history of success since it uses technology to extend the capabilities of the human mind. IA leverages AI and ML to supplement human intelligence while allowing humans to play a crucial role in the decision-making process. 

In addition, AI focuses on confirming or improving work that has already been done, or reacting to events that have already happened. For example, in the investment management space, AI may react to exceptions that have occurred rather than eliminate exceptions from even happening, which could create more work for operations teams. As a result, AI still has not reached critical mass in its ability to solve problems or reduce workloads in the financial services and other industries.

How IA Can Help the Buy Side

The investment management community is in the middle of a technology renaissance driven by shifts in investment strategies that cause greater complexity, and higher trading volumes. In addition, investment managers are facing increased regulatory oversight, investor due diligence and margin compression. By enhancing the work of their human counterparts through an IA-driven approach, AI/ML automation projects have the potential to dramatically improve many buy-side operations processes – particularly data quality and reconciliation.

Many vendors have established experimental labs to identify and address reconciliation use cases for data mapping and normalization, matching and investigations using AI or ML. In each of these use cases, the value of inserting the collective experience of people who have accumulated and contributed a wealth of business and industry knowledge over time cannot be ignored. The question is, how can we capture and leverage that experience to make the best use of AI/ML technology to augment human intelligence rather than replace it?

Take failed trades, for example. Through our experience thus far, we know there is a big difference between AI determining the reason for a failed trade, and IA detecting the likelihood of a trade failing based on a range of conditions and data such as collateral held, securities lending, or stale settlement instructions. The approaches AI and IA take are fundamentally different. Would you rather find the root cause of fails that already happened, or reduce the number of fails and reconciliations? Both are equally important.

Electra’s Approach

* Please note that Gresham has absorbed Electra products. You can now find this solution under Gresham's Control Cloud for investment managers.

Electra has adopted an IA approach that combines the latest AI/ML technology, incorporating human experience and AI learning, with several decades of buy-side operations knowledge and experience in mapping and normalization to optimize the AI component that drives process improvements.

Our AI Research Lab was established to explore industry use cases and pioneer models, statistical and ML tools, computational algorithms and software to address the specific challenges that arise in buy-side markets. The AI Research Lab team combines expertise in core areas such as ML, optimization, data science and algorithms with a deep understanding of financial markets and institutions to make fundamental advances in automation and bring them to the community as market-ready tools.

Here is a summary of the AI/ML capabilities already available today in our reconciliation solution and how they improve the overall reconciliation experience: 

• Data Mapping and Normalization – By utilizing ML to interpret new reconciliation data sources, our solution determines file formats and layouts to automate processing while leveraging our access to bank and broker reconciliation data – more than 1,300 global sources providing over 3,300 unique data feeds. 
• Matching – With more than 100 investment managers reconciling every asset class imaginable using Electra Reconciliation, we combine industry best practices for matching with standard, pre-configured matching logic for a wide range of simple and complex asset and transaction types. 
• Investigations– As the only known solution awarded a patent for intelligently integrating cash, transactions, positions and research information to expedite the investigations process, it automatically suggests the root cause of breaks by accessing relevant data sources.

We expect to discover and explore more use cases for AI/ML that go far beyond identifying a break's root cause. By exposing more and more data to the process, AI and ML will become more of a problem prevention tool rather than a problem detection tool. 

Conclusion

In the investment management industry, where knowledge and experience are highly specialized and complex, AI can go only so far without the human component. With an IA approach to guide the way, all the knowledge gained from years of experience across many individuals and firms can be put into practice to make AI/ML smarter, make a complicated landscape simpler, and maximize the impact and return on investment from AI.