Artificial Intelligence and the language of IR
Every word matters. How many times have you heard that in communication that every word carries weight and to leave nothing to chance? And you might have secretly rolled your eyes and thought ‘here’s another language pedant’. But in the digital age, more than ever, blunt language simply will not cut it.
Today we all know smart machines capture and analyse anything and everything. In what Harvard professor Shoshana Zuboff describes as the age of Surveillance Capitalism, our economic model has shifted dramatically and data is the new currency, analytics the new gold dust. Text analysis (also known as text mining and text analytics) is the process of transforming text into data that machines can understand, process, and interpret, to deliver valuable insights and is used everywhere.
Financial presentations are no exception. If you believed the market practices that have prevailed and served companies well for decades are still reliable, think again—the language of IR is under deep analysis. The very choice of prepositions, verbs or nouns you make, the breaths and intonation you use in your results and earnings statements and presentations to your financial stakeholders can add or subtract to your company market value. Considering that for many, English is not a first language and that often little distinction can be made between homonyms, IR communication needs to tread very carefully.
Harness the power of AI in your financial statements
Today software programmes scan and analyse the tone, words, emotion and underlying implications of the words a CEO or CFO use to present their results. IROs all know the levels of messages that they can disseminate to investors and the reception they are likely to get: in order of best to worst, smart messages that are strategic or tactical impact better than the descriptive or defensive ones. Well, so far, not exactly groundbreaking: you don’t really need the machine to tell you that. But actually, it goes much deeper.
Machine algorithms track and break down so many minute little signals and nuances, from which they compile an assessment that translates into value: what time was the company presentation delivered at? Were questions answered proactively by the management team, or on request? Were statements framed in the past, the present or were they forward-looking? What words were used on the press release? What was the tone, the emotion, the intent? New companies like Alphasense, an amplified intelligence firm that uses ground-breaking AI-powered search tools, helps market players (investors, analysts, companies who wish to stand out from their peer group) “find the money-making needle in the data haystack”. Legacy research tools are too blunt for picking up on value and gaining an edge: AI can harness every piece of company and market collateral (internal announcements, news, filing documents, presentations, research publications….) and tell you whether your latest announcement was better than the last, better than your competitors’ and whether there is an underlying value that isn’t priced into your stock. Others like Q4 Inc. use a powerful AI-engine, iris, to leverage deep capital markets intelligence and use their machine learning capabilities to bring institutional investors and corporate issuers together. “Through this innovative technology, IROs are empowered with accurate predictions around buy-side behaviours and tendencies, as well as the underlying driver impacting these investment decisions, streamlining your targeting efforts and dramatically improving your investor outreach ROI.”, Q4 states.
The Buy-Side lens
Large investment firms also rely on smart solutions to improve wealth management. BlackRock’s famed Aladdin investment technology brings “efficiency and connectivity to institutional investors and wealth managers. It also provides clients with a common language across the investment lifecycle in both public and private assets and enables a culture of risk transparency among users”. State Street’s portfolio management software business, Charles River Development, joined forces in September 2019 with MSCI to integrate MSCI’s portfolio and risk analytics into its investment management system. Big firms, big means, big data.
The algorithms they use contain textual analytics, calibrated to specific industries (announcements from mobility/automotive industry players are very different from that of, say, the hotel industry). Some consultancy companies like Simon Kucher & Partners go further and they develop a calculation app that enables the AI to do a “natural language analysis”, which they transform into a tool that can help companies with pricing power strategies (the Pricing Language Index, or PLI). As such, they have carved out an area of expertise on pricing across all industries –helping companies identify and articulate how pricing power contributes to topline growth, and thereby market value. Boosting an equity story in a crowded landscape competing for funds can earn IROs major kudos. By demonstrating through the PLI how a results presentation was received by the analyst community, for example, the institutional messaging just get sharper and better—and those who ignore the language codes needed to pass the muster of AI do so at their peril.
The language of … love?
Well, some mean feat: transforming words into numbers through an algorithm! In a market where everyone is looking for the silver bullet, is this the panacea? Well maybe not quite: there isn’t always a growth equity story, so focusing on boosting pricing power can fall on deaf ears. And also, the prediction of revenues is often the trickiest part of a CFO’s job, because figuring out one’s market share is often opaque—and the key to value is adding market share profitably in our capitalistic/market-driven enterprises, so the notion of profitability is important, not just top line growth. Another caveat is that analysts are not all equally equipped to quantify the impact of pricing power on the enterprise value, and sometimes because they exist in very different geographies (USA vs Europe for example) and have different ‘inherited’ approaches.
A broader context
In our fourth revolution-impacted world, we know that text analysis tools are everywhere and they help companies make data-based decisions that lead to better products, and services, as well as improved data-driven strategies that respond to customer demands. Part of the resentment many have of Google and Amazon is that they know what you want before you do. Their text and content analysis tools, Google Cloud NLP (train your own Machine Learning model) or Amazon Comprehend (pre-trained NLP models), or others like MonkeyLearn (create custom text analysis models) or MeaningCloud (extract insights from unstructured text data), identify what customers mention most often and how they feel about that. This knowledge is power and it shapes market forces.
The world of finance and capital markets is not immune to this innovation and whilst there is charm and flair about being an ‘old school’ communicator, today’s market requires linguistic acumen, precision and strategic engagement like never before. So, as I said, every word matters.