Sara Benwell: Where is automation standard across the industry at the moment and where do you feel some of the gaps might be?

Markus Ruetimann: There are few industry automation standards as yet. Where they are lacking include factor-based performance analytics, management of unstructured data and classification of emails, to name a few examples.

Our industry talks aspirationally about cognitive technologies and that we should use them to accelerate change, reduce costs and improve capabilities.

The reality, however, is that the augmentation rather than the standardisation of processes is very much at the beginning of its evolution.

Some asset managers seem to struggle articulating operational outcomes and deciding about key priorities.

"Technology-led change is upon us, but across the buy-side industry it will be slower that we currently expect."

When it comes to AI and other cognitive technologies, it is all about prioritisation, simplification and step-by-step execution.

When one wants to automate a particular process, one should explore first whether the process is still needed and, if so, whether it can be simplified before automation.

It is tempting to think that automation resolves the problem, but when you look at a situation in totality, the complexity of dataflows and legacy systems remain the key hurdles to a more efficient operating structure.

Technology-led change is upon us, but across the buy-side industry it will be slower that we currently expect. The sell-side will be asked to make the changes first.

Sara: Where are some of the biggest gaps or where the industry isn’t moving fast enough to keep up when it comes to automation?

Markus: The first area I would highlight is product development, as we all struggle to go through the lifecycle from the idea to bringing the product to market.

Some workflow tools do exist to help manage associated activities such as dataflow design, new operational process, regulatory compliance and the production of marketing and client communication materials. But automation of all the relevant components remains immature.

"We all struggle to go through the lifecycle from the idea to bringing the product to market."

Another area where we all struggle is client onboarding. Again, there are various stages in the process where you have data sets that need to be verified and supplemented.

There are some companies who use natural language processing (NLP) to capture structured and particularly unstructured data. This is a first step to automation.

Sara: What are some of the technologies that are fuelling automation?

Markus: I always think on three levels, operational alpha, incremental alpha and client alpha.

With respect to operational alpha, Software-as-a-Service (SaaS) delivered solutions enable a not-so-technical business to automate processes quickly. There are also products out there which allow almost every employee to be a developer, thus automating processes.

Operating models are changing as the transaction is becoming the atom of future system architectures.

"Software-as-a-Service (SaaS) delivered solutions enable a not-so-technical business to automate processes quickly."

This is evidenced by the rapid growth of near real-time SaaS-delivered data-mastering solutions, such as an investment book-of-record, to surface, control and store data required to make investment decisions.

Incremental alpha is created by using Artificial Intelligence to augment human intelligence. AI is giving us deeper and broader sets of data which help us with portfolio construction.

There are a diverse range of interconnected objectives to portfolio construction, such as correlation, risk, duration etc. and the outcome of the product itself varies from retail to institutional and on the distribution channels etc.

These data-intensive decisions absolutely need AI and I have seen asset managers who use it extensively when it comes to investment research and interaction with their distribution network.

"The final decision making based on the data produced by computers will remain in the hands of humans"

There is of course a lot of debate around whether this creates jobs or threatens them.

I believe that while many functions will no longer require human intervention, the final decision making based on the data produced by computers will remain in the hands of humans. New jobs in data curation and computer science will emerge.

For client alpha, I know of at least two asset managers who have introduced chat bots that are similar to those used for online banking.

If an institutional client is in a different time zone and has a particular question say about their portfolio structure or investment risk, their query is been directed through a chat bot.

If it can’t be answered from the datasets stored, then the query is automatically transferred to a call centre.

The effective use of cognitive technologies is essential to compete in future. But human judgement remains the USP.

What are some of the barriers and risks when it comes to implementing and automating technology?

Markus: There are many barriers from legacy systems to legacy cultures.

There are only a few CEOs in the asset management community that understand the power of modern technology and modern people management.

The former requires investment and access to a skilled talent pool, both internally and sourced from external providers.

"Few CEOs in the asset management community that understand the power of modern technology and modern people management."

Cultures need to adapt so that collaboration and change is encouraged and rewarded.

Employees are becoming stakeholders and the next generation is already technology-savvy and motivated by different factors beyond remuneration.

 

Please Sign In or Register to leave a Comment.