Survey – Use of AI in Asset Management 2018

Shoreline recently conducted a survey of Asset Managers and Asset Owners across Asia Pacific to understand current and planned use of Artificial Intelligence.

The intent of the survey was to answer three key questions:

1) What Asset Managers and Owners within the region are using or planning to use AI?

2) How is it being used?

3) What benefits are being realised

Over the past several years there has been significant and increasing attention given to the role of Artificial Intelligence in transforming most industries. The Asset Management industry is particularly suited to application of AI given its quantitative and data intense nature.
Although there are a handful of well publicised cases of large asset managers currently using AI (Man Group, Bridgewater, Renaissance), we at Shoreline were interested to collect data across a range of organisations to check the actual use of and attitude towards AI relative to the mostly vendor and consultant driven reports in the media.

What is AI?

For the purposes of our survey we have kept the definition simple – Artificial Intelligence (AI) are systems that can perform tasks that typically require human intelligence. See the appendix for a more in-depth definition of the various types of AI.

Headline results

  • 56% of respondents believe AI will be transformational or a competitive necessity for their firms within 5 years
  • 21% of respondents are using AI
  • 60% of AI users are applying it within investment management activities
  • 70% of users are not yet able to quantify the benefits of their use of AI

Respondent profile

The survey was sent to asset managers and owners across Asia, Australia and New Zealand. Thirty-Eight (38) companies responded and represent a cross-section of the industry.

Seventy nine percent (79%) were asset managers and twenty one percent (21%) asset owners. Asset manager respondents ranged from global players down to boutiques. Most asset owners were mid-sized superannuation and pension funds with a few large sovereign funds also participating. The diagram below highlights the spread of FUM of respondents.


The big picture

We asked respondents what role artificial intelligence would play in their business over the coming 5 years. Will AI transform their business or not play role? As indicated by the diagram below, the majority felt the impact of AI would be more toward the transformational end of the spectrum. More than half (56%) believe AI will at least be a necessity to stay competitive – an important finding that I will comment on later.


It is interesting to note that the respondents that indicated AI would have limited or no impact on their firms were all small asset managers (<$10bn FUM), although there were small firms at the transformational end as well.

Who is using AI

Twenty one percent (21%) of respondents are currently using some form of AI. Of those using AI, 90% are asset managers.

Half (50%) have just started using AI (within the last year), 30% have been using it for the last 1 to 3 years, and 20% have been using AI for more than 3 years.

As expected, the big end of town are the primary users – those with FUM > $150bn.

There are, however, small firms using AI. This was surprising given our experience with boutique asset managers and their generally judicious attitude to investment (of both time and money) in technology.


Types and areas within the business where AI is being applied

We asked respondent using AI which types they were using, as indicated in the diagram below.


All the different types of AI (as defined in the Appendix) are being applied by AI users. And interestingly all AI users are applying more than one of the types.

Machine Learning is most commonly used, with Intelligent / Robotic Process Automation a close second.

In terms of functional areas of use, there is a broad spread of business units where AI is being applied – see the graph below.

Most respondents are solely using AI in the front office for investment management related activities. This makes sense given Machine Learning is the most-used type of AI and, in our experience, the most relevant type of AI to apply to existing investment processes using existing / available data sets.

Use in Client Service is also reasonably high with likely application of Chat Bots and Natural language processing to assist in servicing clients. We have seen this use case in a few of our clients, but data quality issues are a key barrier to deployment of this type of AI in a fully automated production application (e.g. directly interacting with clients rather than assisting a client services staff member). Unfortunately, the adage ‘garbage in, garbage out’ still applies in the AI world.


Benefits for those using AI

Thirty percent (30%) of those using AI indicated they were deriving value from its use.


The main benefit stated is enhanced investment decisions followed closely by operational efficiency improvements. Some (20%) are linking AI use to improved investment performance.


Interestingly, seventy percent (70%) of current AI users have indicated they are not yet able to quantify the benefit of AI. Certainly, some of these respondents have not been using AI long enough to quantify the benefits. Based on our experience, other likely reasons benefits are not emerging include incorrect application of AI methods / approaches and selecting problems not suitable to solve with AI.

Like any tool, you need to use AI to solve problems that it is suited for. An understanding of the problem, the AI types / algorithms available and appropriate for the problem, and importantly the type of data required by the specific algorithm to deliver an outcome are fundamental to getting value from AI.

Further investment plans for those using AI

All AI users are planning additional investment in the next 12 months. The emphasis on usage in Investment Management will continue, with other functions receiving relatively the same attention as per current usage.


What about the rest – those NOT using AI?

A substantial portion (80%) of those respondents not currently using AI are considering its use. And like current users, those considering AI usage are firmly focused on applying AI to Investment Management problems (83%).


A range of front office use cases are planned as indicated by the graph below.

The top two responses align to recent conversations we have had with several of our clients who are looking to augment their stock screening process to enable exploration of the investable universe in an efficient manner, and testing for bias in current processes by introducing an unbiased AI-driven voice.


We asked those considering AI usage what benefits they expect to gain – summarised in the graph below


The promise and draw of AI for those considering its use is, in fact, aligned with the benefits current users are gaining – enhanced investment insights and improved investment performance.

The rest

Not everyone is jumping on the AI bandwagon. There is a small percentage of respondents that are not using or even considering AI. Not surprisingly, they are all boutique asset managers with <$10bn FUM –the same cohort that believe AI will not have an impact on their business over the next 5 years.

Two reasons are given for not even considering AI usage:

• Its not currently a priority (67%)
• They are unclear on how to apply it or where to start on the AI journey (33%)

We expected the second reason and, in fact, is exactly why we have established our AI practice – to assist our clients start and accelerate their AI journey.

Concluding thoughts

The most interesting insight drawn from the survey is the disconnect between the respondents view of the impact AI will have on their organisations (and indeed the industry) and lack of response for most respondents. Although most respondents believe AI will have a significant impact on their business and its competitiveness, many have not yet started on the challenging journey to build required capability.

We agree with the majority and believe AI will be transformational – and that transformation will happen quickly as seen in recent history with other disruptive technologies. It is not a matter of ‘if’ but ‘when’ use of these technologies will be essential for asset managers to remain competitive.

Like many technologies, the complexity and difficulty in application is not the technology itself, but rather the cultural and organisational shifts that must occur to accept and accelerate its application and associated benefits realisation.

Acceptance of and willingness of investment professionals to apply AI, understanding the most appropriate way to augment or in fact reshape existing investment strategies and processes, and sourcing the breadth of quality data required for success are, in our view, the key challenges to effective application of AI.

The organisations that will win / most benefit from AI are those that have started building an AI capability within their organisations – incrementally overcoming the key issues mentioned above.

John DiBiase
Managing Director

About the Shoreline AI practice

We have established an AI practice to assist our clients on the journey of exploring and building capability in the application of Artificial Intelligence technologies and methods suitable for their circumstance.AIBlog AIPyramid

Our cross section of knowledge in AI, Investment Management, Operations, and Data and Analytics provides the necessary experience required to work with asset managers and owners to identify suitable problems, develop, and operationalise AI tools.

We focus on results. We support you with implementable solutions and provide you with straightforward access to all our knowledge. By focusing on value creation and minimizing unnecessary overhead, we ensure an optimal return on investment to you.

Appendix – Types of AI

Artificial Intelligence (AI) are systems that can perform tasks that typically require human intelligence. We categorised AI into the most common types / applications:

  • Machine Learning – is a technique to give computer systems the ability to “learn” (e.g., progressively improve performance on a specific task) based on data, without being explicitly programmed.
  • Predictive analytics – encompasses a variety of statistical and Machine Learning techniques including predictive modelling and data mining that analyse current and historical facts to make predictions about future or otherwise unknown events.
  • Natural Language Processing, e.g. Voice Recognition – use of artificial intelligence to effectively and efficiently process large amounts of human (natural) language data.
  • Intelligent Conversation Automation / Chat Bots – is an AI which conducts a conversation via auditory or textual methods. Chat bots use Natural Language Processing to ‘understand’ a request and respond conversationally as a human would.
  • Deep Learning – a class of machine learning whose algorithms mimic the human brain, learning multiple levels of representations that correspond to various levels of abstraction.
    Intelligent / Robotic Process Automation – combines fundamental process redesign with robotic process automation and machine learning.
  • Sentiment Analysis – use of Natural Language Processing and other techniques identify, extract, quantify, and study affective states and subjective information, for example to assess sentiment towards a company based on social media channels.

Share This

Copy Link to Clipboard