Get beyond the hype around AI for asset managers

Updated on

Most asset managers believe data and AI will be key to unlocking business value, yet many still see AI as a superficial solution

The hype around artificial intelligence (AI) in financial services is at a fever pitch. Walking through an airport, bookshelves are stocked with shock value titles like “Our Final Invention”, headlines proclaim “Robots could wipe out 1.3 million Wall St jobs in the next ten years,” and forecast over one trillion dollars in cost savings for financial service companies from AI applications by 2023 (Business Insider). No surprise then that companies are placing big bets on AI, with 2.5x growth in spending on AI systems by 2023 to $98B annually (IDC 2019).

Get The Full Series in PDF

Get the entire 10-part series on Charlie Munger in PDF. Save it to your desktop, read it on your tablet, or email to your colleagues.

Q4 2019 hedge fund letters, conferences and more

Is this your reality? As with any hyped technology, there's a gap between expectations and realizing returns. Many asset managers have identified use cases for AI, but the execution remains unclear. Like teenage sex, everyone is talking about AI, but very few are doing it. Indeed, 53% of firms say they only have or are planning a “limited” number of AI initiatives, according to a survey we conducted.

Opportunities enabled by AI exist across the asset management value chain, yet on the near term we recommend focusing on two in particular:

Marketing intelligence

  • Consider the daunting routine of an investment analyst. From news analysis to scouring through SEC filings, the time-consuming investment research process makes it a prime candidate for the application of AI. Already some firms are changing analyst workflows through the power of automation. For example, machine learning (ML) and natural language processing (NLP) can give analysts the ability to quickly parse through both historical and new digital data sets by layering on social media, image, spatial and news data to predict a company’s revenue growth, its strength compared to competitors and its ability to pay off debt. AI can also automate behavioral signals like geolocation data on foot traffic or online consumption data to uncover trends in supply and demand.
  • One research analyst told us he can now look at forty companies per week using ML and NLP instead of just four or five per week by manually digging through filings and other reports.
  • Now imagine taking that a step further with semantic knowledge retrieval. An easy way to think about semantic knowledge retrieval is if every research analyst had an imaginary friend who could accurately answer any question asked. For example, ‘What is the impact of interest rates on the tech sector in the future?’ This “friend” can look at all of the historical data and then create a prediction using machine learning to come back with a response about the future. What’s exciting about this is the unexpected patterns and opportunities that are revealed that not only help increase analyst productivity but their insights as well.

Customer segmentation

  • Your customers have become accustomed to very personal experiences thanks to customer-obsessed retail giants like Amazon. Asset managers taking a similar path will have a leg up. AI offers this opportunity. Working with one firm, we combined user profiles with transactional information and user behavior from mobile and web usage to create user segments with machine learning for much more precise targeting. However, for this targeting to work, we also needed to segment content, meaning articles like thought leadership and products and offers to cross-sell or upsell. Once segmented, it is easier to match the users to the content. This new approach helped the firm improve its offer acceptance rate from 33 percent to 76 percent. This was not only valuable for the company in terms of incremental revenue, but also provided a better customer experience and more engaging experience as an investor.

Three steps to winning with AI:

  1. Get the data right first. Machines, not humans are creating the logic based on the data they are given. Data is the input, not the output, and given the growing availability and volume of data involved, it is critical to plan for the quality control and data management required.
  2. Adopt a platform mindset. Asset management technologists are used for building applications. If you take the time to build a platform first, you can run experiments, build models and deploy in weeks instead of months with less provisioning time and engineering cost.
  3. Start small and demonstrate ROI. Focus on the types of delivery that are less complex and offer value today like machine learning and robotic process automation. Ultimately, getting this start will help align the company and enable more ambitious use cases with longer timeframes in your transformation journey.

Leave a Comment