Today’s large fund administrators service hundreds to thousands of fund managers and institutional investors. Automation is driving down margins across the investing industry. Price competition has become fierce.
There is a great deal of talk about technology being an edge. And executives are turning to automate repetitive and time-consuming tasks. The goal is to reduce their cost in servicing middle and back-office processing.
Data Ops in Fund Administration Require a Human Touch
It feels like a nightmare scenario for those in the middle and back offices. It’s scary for those who fear the value of their experience and soft skills is not appreciated. Will they be rendered obsolete, at least in the firm’s mind, by software algorithms.
And if the firm pursues mindless automation, the fears are right. Automation that attempts to put aside the knowledge capital in a firm fails.
Change management and growth loops
It doesn’t have to be that way. It shouldn’t be that way. Forward-thinking industry leaders realize that both nuanced human experience and automation processes need to be looped together to get the full value out of both.
Sure, that may seem contrary to most people’s concept of automation in business, specifically in number-heavy industries like fund administration. But those who look to raise profit margins by replacing skilled people with the software will soon find those margins diminishing.
I’ve seen the axiom “smart money does not invest in witless AI,” shared by at least one asset manager in the space.
Applying an 80/20 rule to the matter, what organizations need to understand is that machines do well with predictable situations. You can train machines and program software based upon rules and expected parameters. But machines don’t necessarily understand the complexities of financial managers and how they process their funds.
The concept here is to let the machines do what they do best while letting humans do what they do best, as they work to complement each other. Rather than eliminating the need for human brainpower, machines learn best from people.
Management styles
Jacques Bughin and Eric Hazan have conducted research on this. They see AI implementation working, for instance — when executives plan to grow rather than cut. They invest in technical capability, but also new managerial capabilities to guide change management.
Successful executives commit to getting digital transformation right. But they’re also open to revising their strategic goals as they move through the process.
While they may be flexible on the strategic goals, they are inflexible in requiring rigorous and high-quality data operations. Better data operations make for a better foundation for fund administration.
They also nurture AI ecosystems, which is another way to say they don’t just deliver orders or extract value. They develop communities where members share, celebrate achievements, and back each other up when inevitable challenges crop up.
This is a good general framework for supporting the right change environment. It’s also supported by work that Google has undertaken. They’ve shown the quality of the human touch in management impacts the quality of the technology outcome. Relational leadership, for instance, works.
Making the loop easier
In the end, the employee-machine relationship needs to be better understood to get digital transformation right for fund administration. There needs to be human governance to address the more complicated cases within each process. But the machine needs to use employees’ time efficiently.
The seasoned employee can spot if the automated spreadsheet looks off.
Then they can work with the software to solve the problem quickly. The key is to make this relationship one that grows. That’s better than the human constantly cleaning up after the machine.
The co-existence of man and machine in the fund administration industry works best with exception-based machine learning. This is a relatively new way of approaching middle and back-office data operations.
It’s superior to the market standard extract translate leverage (ETL) approach. There is a subtle but vital difference between the two.
If done right, the nuance amounts to an exponentially faster turnaround time in statement processing when using exception-based machine learning. Faster output means sharper and tuned in the employees.
So, exception-based machine learning makes employees more motivated to stay on top of the process. Employees are more motivated to see the impact of their work. They continue investing in the growing value data operations can deliver.