“AI” to Manage “Big Data” in a way that all active managers can understand and use.
Artificial Intelligence and the incorporation of “alternative data-sets” into the investment process of active managers, hold seeds of hope for the salvation of active management – if you believe what you read in the press.
As usual, the realities of implementation are more complex. Senior managements are frequently challenged to figure out what they want their “AI” to do. Investment teams struggle to integrate alternative data-sets into established investment processes.
However, there is no doubt that if active managers could use and share the (often vast) information they already have, (“Internal Big Data”), it could result in greater alpha generation, and potentially represent a new narrative to engage with clients and regulators.
The Global Projects Center at Stanford University (which works with large asset owners on investment governance/technology issues), has done pioneering work in the field. In 2015 their paper, Knowledge Management in Asset Management laid out the issues and potential opportunities of solving these problems.
Kite Edge, a London-based Fintech firm, www.kiteedge.co.uk is developing new ontology-based search technologies to help asset managers derive greater value from the vast numbers of documents (both internal and external) that they possess. By making this corpus available and discoverable to disparate investment teams across silos, asset-classes and geographies at the Enterprise level, Kite Edge is working toward solutions to the issues raised by the original Stanford paper.
Kite Edge has partnered with members of the Stanford GPC to create a buy-side initiative that will serve the interests of both asset owners and asset managers – who share a common goal in terms of allowing active investment strategies to maximize their returns.
An upcoming Stanford paper (co-authored with Kite Edge), which builds on the previous work, will make a number of critical points:
- Knowledge Management at the enterprise level needs to be a C-Suite priority in order to succeed.
- KM will be an increasing focus for asset owners in asset manager selection.
- he best analogy for KM, is not a filing cabinet, but an Operating System.
- For asset managers, that common language could be an “Ontology” – a multi-dimensional data-map of relationships that can form the basis of a search process specifically designed to help PMs/Analysts test investment hypothesis.
This is why Kite Edge is running “The Ontology Project” to assist in the development of an “operating system” that can add value to the search functionality of any manager – but can then be customized to reflect specific and disparate investment processes.
Kite Edge has developed a unique Ontology specifically designed for the needs of institutional investors. (Ontologies designed for other verticals, like news for example, are ineffective as search mechanisms over investment research).
The Ontology Project is a forum where asset managers and asset owners can collectively contribute to and peer-review the existing ontologies. www.theontologyproject.org
A further facet of the Ontology Project gives asset managers/owners the opportunity for a “Knowledge Audit”. This is a process in which specialists assess the current information architecture of an asset manager (where the information lives) and the research intensity and information pain-points of various investment teams.
This allows the asset manager to develop comprehensive understanding of their KM requirements to take the next step in creating effective, firm-specific Knowledge Management Solutions.
Knowledge Management will emerge as a key competitive differentiator for active managers.
Regardless of funding mechanism, asset managers will have to demonstrate that their research spending is supporting the investment objectives of their clients.
Complex global managers have significant internal “Big Data.” Hundreds of investment professionals receive and generate hundreds of thousands of internal and external documents every year which aren’t stored centrally or searched effectively. Investment professionals conduct countless conversations and generate enormous amounts of useful data in their day-to-day processes.
If this data could be captured and seamlessly leveraged across the enterprise, it would lead to significant gains in productivity and allow the identification of “thematic communities” across asset classes within the manager – generating a new source of alpha.
Moreover, if managers had the ability to reflect their unique investment process (DNA) in the “search” mechanism, they would have a very effective way to demonstrate that they were deriving maximum value from research by systematically data-mining the vast corpuses of documents/interactions that they had purchased.
For active managers that have been struggling to convince investors that they were adding value, the spectre of MiFID II may yet have a silver lining. It may serve as the catalyst to finally re-invent decades-old research processes that have not kept up with changes in technology.
It may give active managers a completely new technology framework and associated narrative to validate their investment proposition with clients and regulators. Most important, these technologies can leverage the attributes, advantages and potential performance diversification that active investment processes can offer versus passive strategies.