contribution - Alfi

How AI will change the AM industry

Sven Muehlenbrock, partner et head of Lighthouse chez KPMG Luxembourg. (Photo: KPMG)

Sven Muehlenbrock, partner et head of Lighthouse chez KPMG Luxembourg. (Photo: KPMG)

Recent, significant advances in technology are disrupting current business models and creating new opportunities in the asset management industry. Some fear the new technologies, some hype them – but there is not necessarily reason for either reaction.

Put simply, these technologies can be useful and powerful for any person, business, or society if they are used properly and responsibly, or indeed disastrous if they are not. They may involve simple robots for robotic process automation (RPA), or more advanced algorithms for machine intelligence (generally called artificial intelligence or AI) 1, which can be used in machine learning (ML) 2 and its subsets like deep learning, reinforcement learning, or Bayesian and artificial neural networks.

The related use cases in the asset management (AM) industry for these technologies are many: using RPA in various AM processes; applying ML algorithms to early signal identification in social media (for predicting price movements, for example); option trading; identifying macroeconomic trends; liquidity management; churn analysis; intelligent document analysis and comparisons; suspicious transaction monitoring; and others. These applications are exciting, but the challenge is unlocking their value with the result not only of surviving – but thriving.

Like others in the financial services sector, asset managers need to create value and secure long-term success. This means increasing revenues, decreasing costs, and thus increasing profit margins in the ever changing digitally driven business environment. How? Generally, the answer involves a (customer-centric) business strategy that properly embeds a digital strategy.

Integrating a digital strategy into a business strategy

The best practice seems to be to place operations and technology capabilities at the heart of a strategic differentiation. This may mean in practice, for example, appointing COOs/CTOs to the management executive committee and seeking new board members who bring interdisciplinary expertise in operations and technology. Furthermore, creating new digital and data roles at the senior management level and investing in talent development to more effectively leverage digital capabilities will become key organisational tasks.

With the right team, a digital strategy responding to the business’s goals can be made – though indeed the strategies can differ widely between, for example, an asset manager focusing on cost differentiation versus one prioritising product and service quality. Finally, the integrated strategy should guide the primary focus: asset growth, operating efficiency, or profitability.

The digital business model

How can a business model be reshaped to become digital? First, bring the business operations and technology function closer together, essentially gluing them together, to eliminate traditional and artificial boundaries. Obstacles here may crop up on the people or internal governance side. Next, define and articulate joint strategic and project plans, which means combining budgets, sponsors, and teams for new initiatives such as RPA or AI applications (like natural language processing) to automate compliance tasks and governance issues (generating statements, reviewing of fund prospectuses, etc.). Finally, keep a clear focus on the client journey and how digitalisation can enhance it based on the business strategy.

Successfully implementing a digital strategy often requires a two-pronged approach: thinking big and starting small. On the small side, asset managers must have a clear and short-term focus on cost reduction through technology. They might tackle legacy issues by streamlining their current architecture, reduce redundancies in and inconsistencies between applications, and retire systems that are outdated or require heavy investments. In terms of thinking big, asset managers must simultaneously consider and invest in next-generation technology (e.g., private clouds or data lakes), capabilities (e.g., hiring new technology profiles), and operating models (e.g., agile applications, new control framework – see below).

Management control for a digital business model

In discussions about digitalisation, it is often forgotten that changes in the business model require changes in the management control framework – the framework that facilitates the execution of the strategy according to the business model. Aligning the business model to the business objectives is often the responsibility of the management. Having a control framework that is ineffective is, especially for asset managers, simply not an option: getting the new technologies wrong means more than financial loss, but ethical, reputational and brand damage, too.

Furthermore, the changes run deeper than it may seem. It is not a simple question of replacing humans with computers, nor will the new controls resemble the old ones 3. For example, if an error hides within an algorithm (or the data feeding or training the algorithm), it can influence the integrity and fairness of the decision made by the machine. This could include adversarial data or data masking as ground truth.

Of course, the business leaders are on the hook for preserving the reputation of the firm, even if AI models increasingly make decisions that are not understood or in line with corporate policies and values, guidelines, regulations, laws, or the public’s expectations. This is when trust weakens or actually evaporates. A new piece of technology, particularly if it relies on algorithms, should meet minimum requirements such as:

- Quality: are its inputs and development process of high quality?

- Resilience: is its long-term operation optimised for change, risk, and uncertainty?

- Integrity: is its use considered acceptable?

- Effectiveness: does it perform as intended?

Inspiration comes from other industries.
Sven Muehlenbrock

Sven Muehlenbrock ,  partner, head of Lighthouse,  KPMG Luxembourg

And another question to be answered: who is or will be accountable for the results of the algorithm?

Ultimately, trust in new technologies is gained only if thoughtful processes, governance, and controls are implemented right from the beginning. On this, inspiration comes from other industries. Aviation, for example, is tightly governed by regulations and each airline has developed internal best practices, which generates trust. New technologies are understood and proven to improve quality, efficiency and safety – this is what the asset management industry must aim to emulate.

In conclusion, disruptive new technologies create new opportunities, but require a digital strategy to be integrated into the business model. Establishing a proper management control framework is pivotal, as is involving everyone concerned by the changes right from the start – by doing so, the new technologies can be successfully implemented and immediately accepted as a valuable aid to business and individual alike.

1.See the CSSF’s white paper Artificial Intelligence: opportunities, risks and recommendations for the financial sector (December 2018) for an introductory overview of artificial intelligence in the financial sector.

2. Machine learning is regarded as a subset of artificial intelligence and refers to algorithm and statistical methods that computer systems use to perform a specific task effectively. They rely on patterns and inference instead of explicit instructions. It is also often referred to as “statistical learning” to emphasise the relevance of statistics.

3. See KPMG’s publication Controlling AI: The imperative for transparency and explainability (June 2019) and the CSSF’s white paper Artificial Intelligence: opportunities, risks and recommendations for the financial sector (December 2018).

More news on the fund industry in  Paperjam’s Alfi  supplement.