Although conventional automation excels at rule-based processes, in asset management it’s the exceptions that require human judgment or involve non-standard workflows that are more difficult to automate and more costly to manage. We spoke with Harry Powell, Endava’s Global SVP, Data & AI, about where AI is poised to drive the next wave of transformation in asset management, in areas that traditionally rely on human intervention.
What are some of the general challenges (usually requiring human intervention) that affect asset management’s middle and back-office?
There are three key challenges: unstructured data handling, judgment-based decision-making and frequent rule changes. Many processes (trade exceptions, corporate actions, reporting) involve non-standard formats like emails and PDFs. Certain processes also often require contextual evaluations that traditional automation cannot handle. Meanwhile, regulatory and market-driven changes create challenges for static workflows.
Can you take us through the specifics to identify the unique problems in the mid- and back-office?
In trade settlement and reconciliation it’s the challenge of handling exceptions often involving unstructured data, such as emails or free-form text from counterparties. These exceptions may require unique decision-making based on specific circumstances, which is hard to codify and means that even though it is relatively mundane, it needs to be done by humans, and that is an expensive overhead.
In data management the challenge is maintaining accurate market and reference data from diverse sources (e.g., Bloomberg, Reuters) which is prone to discrepancies that require manual resolution, delaying decision-making in the front office.
In compliance we are applying bespoke regulatory rules to datasets. That’s labour-intensive and miscommunications between compliance professionals and developers introduces inefficiencies and risks.
With corporate actions management, events like stock splits and dividend distributions involve unstructured formats and non-standardised workflows, again requiring manual interpretation.
In investor tax reporting we have to comply with dynamic, complex tax laws across jurisdictions, and these require frequent updates to processes. Client reporting often requires customised data presentations involving bespoke requests that need different formats for each customer. Frequent changes in requirements for regulatory reporting complicate workflows and necessitate manual updates.
How can AI help address these types of unique and diverse challenges?
AI has many strengths that make it ideal for solving these types of challenges. Firstly, AI can handle unstructured data with contextual understanding; we all know this as we use generative AI like ChatGPT in just this way. AI can also balance competing priorities through autonomous agent systems, also known as multi-agent AI, which Endava is already implementing to great success in healthcare and insurance settings. These teams of AI agents have roles, use tools, follow workflows and collaborate. Networks of AI chatbots can debate together to figure out hard problems and support decision-making with reflection loops to help ensure accuracy. And finally, AI can automate complex workflows and allow human oversight to ‘keep the human in the loop’, helping build confidence in automated systems.
- In trade settlement and reconciliation, large language models (LLMs) can process unstructured data and can automate decision-making frameworks to address competing priorities when handling exceptions (e.g., accuracy vs timeliness).
- In data management AI can integrate data streams, assess data quality and contextualise data for decision-making (e.g., infer the source, validity and ownership of data), reducing reliance on manual interventions and improving usability.
- In compliance AI can automate rule interpretation, flag potential compliance issues and streamline monitoring processes for adherence to investment mandates like ESG or country-specific regulations.
- In corporate actions AI can categorise and extract data from corporate action documents, manage quantitative calculations and automate communication with shareholders.
- In tax reporting AI can adapt to regulatory changes and interpret ambiguous tax rules, reducing the need for bespoke manual coding.
- In client and regulatory reporting AI can automate report generation, interpret evolving regulatory requirements and facilitate self-service reporting for clients.
What impact will this have on operations?
The adoption of AI in middle- and back-office processes is set to transform operational efficiency and redefine roles within asset management firms. By automating tasks that previously required manual intervention, organisations can significantly reduce overheads and reallocate resources to higher-value activities.
- Faster processing times: AI can streamline workflows, particularly in areas like trade settlement, data management and compliance, reducing delays caused by manual reviews or exception handling.
- Enhanced accuracy: By minimising human errors in data entry, reconciliation and reporting, AI improves the overall quality of operations, ensuring regulatory compliance and client satisfaction.
- Scalability: AI frameworks enable organisations to handle increasing data volumes and complexity without the need for proportional increases in staffing.
- Improved decision-making: AI tools, such as LLMs and autonomous agents, enhance decision-making by providing real-time insights and balancing competing priorities, ensuring optimal outcomes.
- Cost savings: With fewer manual interventions and quicker turnaround times, AI allows firms to achieve significant cost efficiencies, making operations leaner and more agile.
These changes allow our people to focus on strategic, creative and advisory roles, helping to further drive innovation while maintaining operational integrity.
What’s the process for getting started and implementation?
To help with the complexity of AI implementation, Endava has developed frameworks specifically designed to simplify the process. Having started working on multi-agent AI over two years ago, we have perhaps more experience than any other major consultancy in building and deploying enterprise AI applications.
Endava’s frameworks provide a structured and proven approach, enabling clients to adopt AI with minimal disruption to existing systems and processes. The frameworks are adaptable, integrating seamlessly with a client’s operations. We combine our expertise to guide clients through implementation, aligning frameworks with specific requirements.
We can deliver a tailored proof-of-concept in four weeks to demonstrate impact in a short timeframe, allowing organisations to move forward with confidence.
Closing thoughts
Managing trade exceptions in asset management requires navigating multiple stakeholders, complex trade-offs and dynamic decisions. While automated systems can identify some issues and resolve simple exceptions, they lack the contextual understanding and nuanced judgment that humans provide. Human expertise ensures decisions align with fiduciary duties, regulatory standards and the broader interests of all stakeholders.
However, AI’s ability to handle unstructured data, adapt to regulatory changes, manage trade-offs and support balanced decision-making gives a competitive advantage to those businesses that adopt it. As asset managers navigate growing complexity and regulatory demands, AI provides a scalable, future-proof solution that drives operational excellence and opens doors to new growth.
Like to learn more? Vist our AI hub or read more about agentic AI in our blog.