Since its inception in the 1950s, artificial intelligence (AI) has seen at least two major hype cycles and long winters of disillusionment.. While artificial intelligence suffered through the recent disullusionment cycle from the 1990s to today, its facilitating and corollary technologies have thrived, and we are now entering into a new boom in applictions of artificial technology.
Despite the difficulty in defining AI and its subfields of related technologies 1, the buzz about fintech has continued to grow. Financial services have been revolutioned by the computational arms race of the last twenty-plus years, as technologies such as big data analytics, expert systems, neural networks, evolutionary algorithms, machine learning and more have allowed computers to crunch much more varied, diverse, and deep data sets than ever before. Combined with vastly more powerful computers at cheaper prices and the advent of social networks, mobile phones and wearable devices, and financial services companies are entering a new wave of possibility in how to leverage artificial intelligence by enabling computers to watch and learn. 2
While most of the businesses built around machines making decisions aren’t deploying true AI today, they are using data-intensive technologies that will help technologies and companies continue to get closer to implementing AI in commercial applications.
Despite the hype of intelligent machines, the first applications of AI aren’t replacing humans and human intelligence but augmenting them. Text-based conversational chat has been embraced by many startups as a way to deliver a personal assistant-like experience in many industries, and in fintech we’ve seen the example of companies like Kasisto utilizing this UX and pairing it with AI to scale the impact of humans using technology. Instead of being bounded in customer support applications by humans responding to users through chat windows, AI and related technologies are being applied to deliver a human-like chat experience without the need for nearly as many human assistants.
By pairing this user experience with smart agents that can analyze and crunch data about individual behavior and compare to broader datasets, small and big companies could be able to deliver personalized financial services as a scope and scale never possible before. Consumer banking, advisory services, personal financial and investment planning advice, wallet and wealth management, all of these services can be delivered utilizing a conversational UX powered by AI. The mix of technologies can enable companies to provide services to segments of segments of customers where they were unable to provide high-touch, human service profitably (i.e. lower net worth segments for personal financial, investment and retirement planning and advisory), but can now serve using codified knowledge and AI-powered software.
In addition to new segments, they can be more personal, providing advice at the transactional level (i.e. every single transaction), part of the promise of smart wallets like Wallet.ai. Imagine having an assistant wih you to help you consider, analyze, price, and consider every single thing you spend money on, at a granular level that no human assistant could assist you with. Is a roboadvisor that offers you rules-based advice based on a set of predefined parameters AI? Probably not, but newer technologies in the future that are based around watching and learning about your behaviors at the individual lavel, rather than the collective level, could deliver advice and outcomes that are individualized in a way never possible previously.
But in additional to new services that use this kind of UX as a differentiating factor, expect to see larger financial services companies integrate conversational UX powered by AI as a way to provide better, faster, and cheaper customer support to consumer and business customers. Some portion of interactions will be still be handled by humans, but AI can enable companies to deliver better service and refocus consumer support costs on higher-value areas.
The idea of augmenting human interactions and intelligence with AI doesn’t end with consumer-facing products. AI could power technologies that overlay humans to provide an oversight and tracking mechanism to employee actions, helping with compliance, security, and the monitoring of employee actions. Monitoring discrete, repetitive data entry tasks, computers could watch and learn over time to verify data entry and test for specific events, assess risk, and find fraud. Any area of fintech that is regulated creates the opportunity for companies to deploy AI-powered employee and systems oversight.
But the opportunities in the enterprise don’t end with inward-facing applications. Lending and underwriting products could be augmented by AI technologies that allow computers to process data and make decisions faster, better and easier than humans alone. While it’s still to be determined how new data sets created by technologies like wearables and internet of things can be used for credit and insurance decisions, AI-based technologies make it more possible for firms to use these new datasets in highly personal ways at scale. Testing for fraud, risk, and important events is difficult to do at a truly individual level, and as a byproduct it’s defined the business model of many financial services companies, but augmenting human data decisioning is an area with high potential.
Fintech firms have developed and used AI-related technologies for years, creating evolutionary algorithms to help them learn, find, and act on existing and new heterogenous data sets created by popular new technologies such as social media, mobile phones and wearables. But AI is creating larger opportunities to go beyond matching and testing data to create more "intelligent" traders and trading systems, using robotraders to test and optimize predictions and trading rules. AI can help oversee and augment trading decisions and rules, helping process the data and creating the algorithms managing trading rules.
Firms have built trading algorithms based on devining sentiment and insights from social media and other public data sources for years, but technology companies like Dataminr and others are deploying platforms for a larger set of companies to use. Accessing and utilizing large, heterogenous datasets is becoming possible for far more firms to utilize, so how will firms leverage and build on top of these datasets?
Fintech is in an interesting position in that the technologies around artifical intelligence technologies have been tested and deployed in specific applications for the last couple decades and have powered much of the innovation in financial services. While much of the investment in artificial intelligence has been into multi-purpose platforms that are still figuring out their specific, high-value usecases 3, the opportunity in fintech is a bit different. Fintech has a base of technological prowess in the technologies supporting AI and a number of immediate high-value applications.
At first, AI may deployed more in back-end technology settings to power large-scale decisioning in lending, trading and financial analysis, but it could also be a technology that expands how everybody interacts with financial services firms. Many of our complaints with fintech today are around the difficulties in getting to real, quality, personal service, and perhaps it’s an artificially intelligent agent that helps deliver cheaper, faster and better personal services. 4
More reads on the topic: How Long Before Superintelligence?, There Will Be No Line Between Us and Our Devices, Siri’s Inventors Are Building a Radical New AI That Does Anything You Ask, Is Artificial Intelligence the Way Forward for Personal Finance?, The Next Frontier in Banking: Big Data and Artificial Intelligence, Artificial Intelligence and Wall Street Trading, The non-bank of Facebook, AliBay, and AI: The future of global fintech, Banking Startups Adopt New Tools for Lending, Facebook AI Director Yann LeCun on His Quest to Unleash Deep Learning and Make Machines Smarter ↩