FINDALL

Creating Thematic Universes On-The-Fly

An AI-based tool, to find relevant companies for any investment topic

Introduction

We provide this description as an example of how some of our tools can be applied and used by our clients. Should you require additional detail or technical specs, we would feel more comfortable discussing further over a call as some of our techniques are quite proprietary and part of our intellectual property.

What is it?

FINDALL is an application based on NLP technology and proprietary algorithms developed for ALLINDEX clients to generate thematic investment universes at will.

Why is it relevant?

It addresses the pain point of creating thematic universes and maintaining them. Maintaining thematic universes is the reason why asset management companies can’t scale the number of thematic investments. FINDALL not only creates initial thematic investment universes, but also runs periodically in order to manage existing universes, thus putting no limit on the number of thematic universes to be maintained at the same time.

What does it do?

FINDALL applies a two-step process to generate a thematic universe. After clients have put in the thematic, FINDALL creates in a first step a semantic model, which uses the technical terms characterizing the thematic. This semantic model is used to crawl the web and also query a data lake of news articles. In a second step the search results are analyzed by artificial intelligence algorithms such as transformers (e.g. BERT) to perform an NLP technique called NER (named entity recognition). 

The client is presented with a list of companies and the sources where those names have been found. The list can be curated and downloaded for further refinement or analysis.

What makes FINDALL unique?

The application solves two challenges with creating automatically thematic universes: It is general purpose by using a semantic model, which is created ad hoc. Therefore, a list of terms, describing a thematic is not required and there is no need to maintain such a list. Secondly the application is agnostic to the sources containing the information. It can read websites, pdf-files, office documents, pictures etc. without the need to specially analyze the format of the sources. 

The clients have full control over the application and can run as many searches as they want. The results can be curated by the client and then downloaded or integrated into other applications. ALLINDEX’s involvement in the search process is limited to provide the technical infrastructure.

What could be a commercial use case?

Thematic investing is a growing area of interest for investors. FINDALL can reduce the effort to launch and maintain very fast thematic investment ideas, which are then wrapped into an ETF or index linked fund. It offers, however another interesting business development angle, which can be discussed with an ETF provider. In order to reduce the risk of bringing a thematic ETF to the market the thematic investment idea can be presented to clients of a direct indexing / managed account service. This service can either be accessed through a partner or built inside the ETF provider. If the traction of any particular topic is above a certain threshold, then an ETF can be brought to the market.