4 Challenges of Operationalizing AI into Investing
Updated: Jul 8, 2021
This article is part of Seleya's blog series "Investment Research: Time For Innovation".
Many institutional investment platforms have begun adopting data, and in some cases AI, to modernize their investment research processes.
Yet, incorporating data & AI into fundamental investment processes is difficult and there is no industry blueprint to guide the path.
Based on our experience of operationalizing AI into the world’s largest asset managers, we observe four common challenges.
Challenge #1: Organizational Inertia
Modernizing investment research processes is, first and foremost, a change management exercise. While organizational inertia exists in virtually every company, fund managers tend to adopt innovation faster when three enablers are aligned:
Individual investors feel enhanced by their peers or collective wisdom (i.e. "social network effects");
Tools benefit the entire organization not just select teams. This encourages investment dialogue across the firm;
Strategic vision communicated by C-level executives with consistent leadership and execution follow-through.
Challenge #2: Decentralized Efforts
Decentralized pilot projects naturally gravitate towards team-level objectives in search of “quick wins.” This is because of the fragmentation of interests amongst teams with different investment styles and risk appetites. However, this path inevitably leads to low-ROI “toys” that are often not scalable or reusable by other parts of the organization.
Challenge #3: Unrealistic Expectations
Fundamental investors often incorrectly expect data science will generate alpha signals with high efficacy. However, a signals-based mindset is incongruent with concentrated portfolios and a mosaic-based research approach. In addition, quant signals tend to incorporate information that is backwards-looking.
In other words, fundamental investors are best-placed to evaluate the qualitative prospects of a business on a forward-looking basis and cannot be replaced by machines. Human fundamental insight should be at the center and enriched with data perspectives.
Challenge #4: Lack of Integrated Expertise
One final challenge that asset managers face is talent. Specialized AI experts are hard to come by, and there are even fewer industry professionals who have domain expertise integrating investing, data science and AI. The success of operationalizing AI depends on the intimate collaboration between investors and data scientists.
In our next and final blog post in this series on investment research, we will offer Seleya's five guiding principles on operationalizing AI for success. Stay tuned and follow us on LinkedIn / our website at www.seleyatech.com.
Seleya Technologies is an industry expert in AI and quantitative analytical tools for financial institutions. We leverage AI to augment human perspectives, enabling financial institutions to make decisions faster, more accurately, and with less bias.
Our two solutions include ExpertAI for institutional investors and ExpertAI ESG™ that scales up a financial institution’s in-house, proprietary ESG assessments.
Our team has been authoring the intersection of AI and investing since 2013. The company is founded by experienced investors and computer scientists with over 20 years of experience developing solutions for financial institutions.