There was a time when financial data was scarce, orders were collected and executed manually, market data was primarily handled on spreadsheets. That world is gone.

There was a time when financial information was channelled through conventional media and could be consumed only by reading newspapers or watching the news. That world is certainly also gone.

In today’s digital world, decision-making requires individuals and their financial advisors to digest big data and harvest investment insight behind the news, gathering real-time knowledge about unconventional relationships between prices, corporate deals, market changes, economic interests, political shifts and human behaviours.

As markets change faster than ever, insights-based decision-making needs to be made available to front-line investors in a timely and digestible manner. This requires a Copernican change in investment management operations, opening risk-based architectures to real-time data analytics and allowing front-end digital applications to consume and visualize investment analytics that were previously restricted to the ivory towers of risk departments and trading floors. Luckily, recent FinTech innovation is generating a shift in how technology is architected, providing an opportunity to change how the world interacts with financial technology. As data and information is nowadays very fluid, analytics supporting decision-making should also be agile and flexible to design alternative investment experiences (thus added-value workflows) without loss of consistency. Financial technology to date has been architected around specific workflows driven by specific client requests or idiosyncrasies on how they tackled the problem to solve. The creation of robust financial services platforms, hosting a modular set of FinTech microservices, is the solution to the re-engineering task addressing investment decision-making in the digital age.

Microservices are building blocks to assembling workflows. A complex application can be built by stringing together a collection of individual components, instead of attempting to fine-tune a rigid container. Just like during earthquakes, rigid architectures have no chance to adjust to even small changes in the terrain, and thus become fragile and break. Microservices, instead, are flexible API-driven analytics which can be accessed via cloud-based investment platforms and consumed in a fit-for-purpose modality, granting flexibility to enhance decision-making by calling new calculations, dropping outdated sources, mixing a variety of investment signals as markets and sentiment change. Having deconstructed investment analytics into a set of streamlined microservices, and hosting them on a scalable platform, machine learning (ML) techniques can be deployed to exploit this powerful constellation of data-driven, information-based, “insights-as-a-service”. A cognitive process can be layered on top, aiming to tap into the streamlined microservices more easily than a closed-off, opaque software package. This cognitive layer, which consists of ML techniques, is the ultimate game changer, since it augments investors’ intelligence based on actionable insights.

How Could This Work in Practice?

Since investment insights can be modelled and tested as scenarios, a set of scenario-based risk management functions can be built as microservices that more easily architect a solution to a particular business problem. Insights-driven scenario generation tools can model a given shift to a particular risk factor. Cognitive computing expands the hard link between risk factors and portfolio holdings. The source of risk is no longer the risk factor itself, but the point in time and short-lived “insights factor” which can be sourced by a machine learning process that links back what happens outside a portfolio to a potential change in investment performance. The risk factor is nothing else but a connector between cognitive insights and final investments. Clearly, as modern data/information systems become ever more complex, digital visualization is key to create enough transparency to act upon given cognitive insights. “Knowledge graphs” are therefore essential to leverage data on companies, their competitors, their key employees, their supply chain, their market of reference and potential contagion from apparently unrelated instances. Natural language processing can break down a news article into its entities, relationships and concepts. Determining their relevance becomes simply a matter of finding a linkage between an article and a knowledge graph.

Cognitive-enabled architectures based on “insights-as-a-service” microservices and scenario analysis have deeper ramifications for the entire industry at large. The complexity can be abstracted away, such that any financial advisor has the ability to address client concerns like “what does this news article mean for my investment portfolio?” or “what exactly am I invested in?” with a single mouse click. This provides clients with real value: the information they need to more confidently make investment decisions. That is, in essence, digitization of knowledge.

Why Knowledge Digitization Will be a Norm in Investment Management

Financial markets have experienced a structural change since the 1950s, when modern portfolio theory (MPT) was initially formulated. Until the financial deregulation of the early 1980s, markets were fairly simple and financial products were fairly simple (the first fixed-income index appeared only in the 1970s), thus MPT provided a tale for a simple world based on mean-variance assumptions. MPT assumes a world of investment symmetries, where one-size-fits-all optimal portfolios could generate unprecedented wealth for American households of different risk profiles. With the 1980s, the Western world entered a more asymmetrical phase: markets became highly volatile and financial products became highly complex (for example, structured finance, subprime notes). This allowed financial institutions significant power to trade the asymmetry of information and develop market views as the main wealth management narrative, which continued until the global financial crisis (GFC). In the aftermath of the GFC, regulation stepped up to impose higher transparency in investment decision-making and reduce the information asymmetry of financial markets. Markets are still very complex, financial products are getting progressively simplified (see the rise of passive investing) and transparency is the new mantra. Broader MPT criticism suggests that optimal portfolios, as professionals have known them for decades, might not truly exist. In a world of higher transparency about risks and costs (thus reduced asymmetry), low interest rates and margin compression, unusual correlations among risk factors, faster than ever propagation of market-sensitive information and the focus of investment managers is shifting towards the generation of real and perceived added value for final clients. That means assisting investors and advisors to optimize their decision-making by providing intuitive insights into the risks and uncertainties of financial markets. The scenario-based contextualization is centred on cognitive dialogues of how news and events can transform into scenarios, and thus potentially impact investment performance. This becomes a powerful attribute to facilitate an engagement mechanism based on better understanding of financial risks, and facilitate the building of sounder decision-making processes based on “insights-as-a-service”.