Every so often, I hear of something going on in the market that piques my curiosity and that I feel I should understand better. Over my IR career, developments of this nature have often had an acronym associated with them – OTCs, xbrl, NEO, ESG. Most recently, it has been AI (Artificial Intelligence) and its less-known associated acronyms NLP (Natural Language Processing) and ML (Machine Learning).

Sometimes my research convinced me that the phenomenon behind the acronym in question was good to understand but of little relevance. This was not true of ESG – which now feels ubiquitous in the IR vocabulary. Nor will it be true, I suspect, of AI – although it is still early days.

We all likely have heard of AI applications within our businesses and consultants like McKinsey are writing extensively about the topic[1]. In my company’s case, we are working with the AI community to help us optimize planograms and merchandising, leveraging data around customer behaviour patterns. And if your company, like mine, is using some of the more advanced analytics features of Microsoft, you may have been surprised in the recent past to have an automated email generated telling you how many collaborators you work with or which document you might need for that 2 p.m. meeting. And, of course, pre-COVID roadshows had already become much easier with insights telling you how long it will take to get to your next meeting and how to get there.

But on an IR-specific level, there are some signs that AI is becoming more relevant, as the buy-side adoption of AI tools picks up.  

The use of data and algorithms to parse big data is not new to investors. Algorithmic trading by hedge funds and in the index/ETF world has been around for a while. But as a number of publications by firms like PWC[2] and BlackRock[3] have suggested in recent years, the use of AI is now penetrating beyond improving trade execution to the generation of investment ideas.

AI applications ultimately depend on the existence of huge datasets. One of the limitations traditionally cited around the use of AI/ML/NLP tools was data quality. But in recent years, our company documents have moved into more searchable online formats, databases and tags and transcripts for events held virtually have made information more accessible, and other less traditional data is becoming available for purchase. Today there is more data and more powerful tools being developed to search that data.

As the BlackRock report points out, “there is now a broad range of [information beyond company filings] that can signal a company’s future performance. Asset managers have developed AI and ML tools to compile, cleanse, and analyze the universe of data available, including analyst reports, macroeconomic data (e.g. GDP growth, unemployment) as well as newer ‘alternative’ data sources. Examples of alternative data include GPS and satellite imagery to see where consumers are going, internet searches and tweets to see what people are researching and talking about, and employee satisfaction data, all of which can be accessed online today. These data points can help portfolio managers better assess individual companies and sectoral trends. As the volume of real-time data available increases beyond the capacity of individuals to analyze and understand it, the ability to compile, cleanse, and evaluate that data is increasingly important. AI and ML enable asset managers to find patterns in this data at scale, potentially identifying signals for generating returns for clients.”

And on the ESG front, the desire to automate part of the research and screening process for index passive investors appears to be part of the reason for their push for common standards and frameworks, allowing them to more readily identify patterns and generate returns.  

Looking to the U.S., NIRI’s 2020 Think Tank Report on the use of AI in Investor Relations observed that “AI use in IR is in some ways at a nascent stage,” with “most IR professionals [not being] heavy users of AI-powered tools in their internal day-to-day work.” However, like the BlackRock piece, NIRI noted the use of AI by IR’s external audiences and suggested that IR professionals could be at a disadvantage if they did not “quickly come up to speed to understand AI and its implications.”[4]

As one retail sector IRO noted to me in an interview for this piece, “investors leveraging data and satellite imagery of trucks at distribution centres” may eventually come up with enough real time insights to make last quarter’s earnings report seem a little ‘stale’ by the time it is communicated. On a personal level, I have recently had incoming contact from a start-up company building an analysis tool for a Canadian investor leveraging publicly available data to try and identify likely future climate transition risks, which included a view of likely impacts on financial performance.

While the widespread development and adoption of more sophisticated tools will probably depend on investor resources and appetite (and doesn’t threaten to make the earnings call obsolete just yet), there are also some tools in use today that IROs should be aware of and consider in their work. For instance:

• Generation of earnings headlines by automated journalism is pulling data from filings as well as releases and driving a news agenda (and trading) that can be at odds with the sell-side sentiment and that of active equity managers. 
• The ability to leverage natural language processing for analysis of earnings calls and transcripts is increasingly being used by investors to seek out key themes and to analyze sentiment, suggesting that IROs would be well-served to think even more than they did historically about the use (or overuse) of certain words or terms.
• The types of AI-informed tools on the market today are also generating invaluable insights and efficiencies for IROs using them to inform their competitive analysis, speeding up time-consuming peer transcript and filings analysis and generating summaries of peer commentary around key trends.

As one IRO said to me, “when I think about the opportunities the tools are providing me to focus on the high-value-added activities, my mind turns to the other mundane or time-consuming tasks that AI might eventually automate – imagine how much opportunity there could be to inform things like targeting or maybe even the first draft of the CFO’s comments for the quarter.”

Imagine it for now. It feels like there is a very real chance that it could be in our future.