The past years have been great for followers of artificial intelligence (AI). We’ve seen governments implement pioneering systems of surveillance and law enforcement, constantly tracking citizens everywhere they go. DeepMind unveiled a system that taught itself how to beat everyone else at chess—in less than 24 hours! And if you haven’t been living under a rock, you will know Tesla has set self-driving vehicles loose on public roads.
Financial services in transition
It is proving to be a fascinating time, with users and businesses alike experiencing a variety of milestone moments in AI and IA (Intelligent Automation). You might even call it a new era. While the concept of artificial intelligence is certainly nothing new—it has been around since the 1950s—its development has been slow and laborious until now. But as recent developments accelerate these technologies from their theoretical beginnings into real-life applications, AI has emerged as one of the hottest topics of modern business.
We’ve identified three key levels of implementation: Strategy, Service and Stream.
Following an explosion of opinions and essays exploring the new opportunities and challenges that AI brings, it became clear that the wide and complex field needed to be structured. In the end, we’ve identified three key levels of implementation: Strategy, Service and Stream. In this article, we look back at the year’s landmark moments in terms of the three S-levels of implementation. What were the highlights, what were the challenges, and what can we learn from it all?
What are 'the 3 S-Levels of Implementation'?
- To ensure competitiveness and market relevance, financial services providers have set about creating a sustainable AI Strategy.
- AI-powered solutions are capable of delivering new Services to customers.
- With the maturation of the artificial intelligence industry, policymakers are more than ever working on frameworks and pushing companies to avoid ethical and legal issues (Stream).
Strategy Trend: Progress through partnership
It is beyond doubt that AI is making a huge impact on the financial industry. Almost three out of four financial services executives believe that, by 2020, artificial intelligence will have completely or significantly changed their organization.
Correspondingly, 2018 was an excellent year for FinTech funding. Worldwide developments were seen, for instance in the United Kingdom with ‘neobank’ Revolut, in the United States with ‘InsurTech’ Root, cryptocurrency exchange operator Coinbase, and payday-loan FinTech Earnin, and in Hong Kong with alt lender Oriente.
The funding rounds held by these FinTechs raised over $100 million in 2018. Since 2010, over $50 billion has flowed into global FinTech start-ups. With its agility, cost-effectiveness and fast time-to-market, financial service providers are more willing than ever to collaborate with FinTechs. And collaborate they do – if these companies don’t have a queue of FinTechs knocking at the door, it’s probably because they’ve already welcomed them in with open arms.
While FinTechs are actively pushing innovative AI-powered solutions to the market, outdated regulations could be a hindering factor for adoption and scale-up of operations. Therefore, one of the biggest challenges is to develop FinTech sandbox environments. The sandbox creates a ‘safe environment’ for FinTech entrepreneurs to test their new technology-based products—it’s a live system that can be accessed for a limited time, without the necessity of full authorization and licensing processes that would otherwise be required.
Several governments have responded actively to give FinTech the space it needs to grow. A good example is the Hong Kong Monetary Authority (HKMA), which launched the FinTech Supervisory Sandbox (FSS) in 2016. Two years later, it has funded 29 pilot projects in the areas of biometric authentication, soft token authentication, chatbots and distributed ledgering. Governments of countries like the United Kingdom, India, Kenya, and Dubai are also at the forefront of developing FinTech sandboxes, while Dutch financial supervisors have also touched upon clearing the way for regulatory sandboxes. If you work in financial service provision or a dependent industry, now is the crucial moment to start looking more closely at FinTech developments – or risk being left behind completely.
Service Trend: AI gets soft skills
In May 2018, Google unveiled and demonstrated its latest digital service, based on its Duplex technology – a vocal robotic assistant that can make calls and book appointments. What it does is not only useful, but the way it does it is revolutionary: Google has created a bot that can speak and converse in a natural way, like a real human being. For many working in AI, successful natural language processing (NLP) is akin to the holy grail.
In 1951, British mathematician Alan Turing introduced a way of testing whether a machine can mimic an intelligent human being. This so-called ‘Imitation Game’, later named Turing Test, has received significant attention in 2018, looking at AI-based services—not least because Google’s CEO claimed their new assistant had ‘passed’ the test. If it did pass, expert opinions differ, it is a world first. In any case, it is a fantastic achievement and a huge step forward.
Following this major development, voice assistants are expected to fuel big changes in service delivery and customer experiences. A recent survey among marketeers in the US showed that 87 percent of the respondents believe that voice assistants will have transformed customer experiences by 2020. Until now, most applications of voice assistants in traditional services are experimental and limited to specific tasks. In the next few years, voice assistants are expected to revolutionize the interaction between business and consumer by using multi-user interface service models to offer more personalization and higher-quality services.
Stream Trend: “And if your AI told you to jump off a cliff - would you do it?”
The implementation of artificial intelligence strategies and services is progressing more quickly than ever before. It seems that, for many industries, now is the time to really start thinking about how to adapt and benefit from intelligent technologies. However, before AI solutions can be successfully implemented and capitalized upon, the business impact must be acknowledged. You’re probably wondering what that means in real terms - let’s take Amazon’s recruitment tool as a starting point to discuss what acknowledging AI means and how you can do it.
Following a successful implementation of machine learning in its e-commerce strategy, Amazon started using AI to assess recruitment candidates in 2014. Unfortunately, it didn’t work as planned. Operating in a male-dominated industry, real-life historical training data showed that most positions were occupied by men. Consequently, Amazon’s tool preferentially selected male candidates over female candidates. After attempts to neutralize the algorithms and remove the gender bias, Amazon lost trust in the project and pulled the plug.
Amazon’s problems are an example of the challenges that come with stream-level implementation. An AI system will produce results – but how do you know if it’s really working as intended? ‘Responsible AI’ and ‘Explainable AI’ became more important than ever in 2018. They should be a crucial consideration when developing and training algorithms.
The fact is, an AI let loose on the world without proper management could be disastrous… With new technologies handling customers’ money, assets and sensitive data, it is easy to imagine the backlash if something goes awry. Complex technological developments are happening at a rapid pace and customers are increasingly empowered. As a result, Stream-level considerations are now a base requirement for creating a competitive technological advantage. It is essential for any financial services provider to have Stream-level management at the core of its automated service portfolio.
Over the next few years, the financial industry is expected to make big developments in Stream-level implementation. From client account management to debt profiling, there are potential game-changing solutions around every corner. But be warned - these technologies are not ready to be used with blind trust, straight out-of-the-box. Instead, now is the time for careful trials and calculated risks.
Spread your wings and (butter)fly away
In today’s AI-based services landscape, banks face increasingly fierce and innovative competition. We have yet to see how the competitive landscape will transform in 2019 and how future customer needs will be served. While it is important to adapt to the changing landscape of financial services, technological development is only part of the whole picture. To become or remain competitive in offering these solutions, it is paramount to ensure AI systems make responsible, explainable decisions.
More interesting developments will emerge this year, building on the foundations laid in 2018. The primary question remains: how will the competitive landscape transform in 2019? Will future customers be served with a totally new approach? Will FinTechs gain more market share and generate a ‘Butterfly Effect’, significantly impacting the competitive landscape? Or will banks leave their cocoon and transform from a caterpillar into a butterfly by incorporating the 3S levels in their core? Whatever the path, it is clear that AI will have a significant impact on the financial services industry. It’s up to you to decide where your wings will take you.