People are the path to AI, just ask them
21 January 2020
In the second piece of a series on the future of technology, Pirum Systems and Patrick McCreesh break down some of the most prevalent myths around the consequences of bringing AI and automation into financial markets
Image: Shutterstock
‘Dear tech, let’s talk’ is the sentimental opening to a recent IBM advertising campaign. This anthropomorphism of technology is not just a clever quip; it raises interesting and provocative questions about how we view and interact with technology in the workplace today. It is clear that technology will always be changing how we work, and we should be encouraged to engage in conversations about this fact.
Over the past century, technology has vastly distorted and disrupted the landscape of the workplace. From Henry Ford’s assembly line of the 1910s to Toyota’s ‘just-in-time manufacturing’ from the 1970s and, more recently, online help bots for sales and customer support as firms have strived to reduce human-based procedural overhead to add value in more important facets of business. Perhaps the most notorious disruption to the workplace in the past decade has been the introduction of achievable and effective automation and artificial intelligence (AI).
Automation and AI are coming (if they haven’t already arrived) to the financial securities market and will vastly change the way teams function. We are already beginning to see these technologies being integrated into the modern work environment, and this mass propagation has introduced a provocative tension between employees, stakeholders and the technology itself. The automation and AI market is set to reach $70 billion in 2020, according to Forbes, and it is therefore paramount that firms address the origin of the tension felt by their employees which is borne predominantly from a misguided fear of human redundancy.
It is no secret, then, that the cleanest process for integrating automation and AI into critical aspects of business is acknowledging that it is not the technology at the heart of successful AI, but rather the people. By engaging your teams on how you intend to leverage the power of Automation and AI to lift the shackles of cognitive fatigue and stress associated with menial, repetitive tasks, you create an atmosphere of optimism over adversity. Employees are less likely to make disastrous mistakes in data entry and rationalisation and can instead begin to focus on adding value to the business in other, more creative ways.
Due to the manual nature of many processes, the securities finance industry has not been alone in encountering challenges relating to human input errors. Firms often encounter erroneous bookings made by trading desks and persistent idiosyncrasies surrounding static data which take a significant amount of time to understand, let alone, remedy. These can often result in detrimental outcomes for market participants and even the smallest error can leave businesses dangerously exposed down-stream.
Using tools from the modern deep learning and automation arsenal, we can introduce trust-worthy, data-driven solutions to these problems that will engender a symbiotic relationship between AI, automation and employees. This will boast huge advantages such as boosting productivity, creativity and morale. Because AI predominantly excels in pattern matching, it seems that the securities finance industry could be an optimal ‘playground’ for these algorithms. We are confident that in the near future solutions will begin to provide ‘what if’ scenarios and simulations with narrowing latent space to provide collateral management efficacy, among other post-trade processing benefits.
The IBM advertisement strikes at the heart of what needs to be acknowledged – namely that we have an inescapable and intrinsic relationship with technology in our modern lives. This intrinsic relationship aptly justifies the bold use of personification. Technology, like a ‘dear friend’ whom we may indeed write a letter to, helps us complete goals, organise ourselves and connect with the world on the international playing field. In recent years, technology has garnered good press for its ability to connect the world, allow social mobility and simplify our lives. However, current discourse surrounding automation and AI does not reverberate with the same levels of optimism. Instead, a mention of AI or automation is shrouded in numerous layers of mystery and fear of almost mythical proportions. Why is our potential relationship with AI so different to that of other technological advancements in the workplace?
Myth #1: AI is just automation with extra steps
It is extremely important to educate employees on the nature of AI in order to help direct thought processes about the new technology.
Perhaps the most important distinction to make between AI and automation is that of routinisation. Automation is the adoption of an approach that removes routine work from the process flow of organisations. The integration of a factory-like, blue-collar style to white-collar work has created aspects of our jobs that are repetitive and can be streamlined through automation. In order to deploy automation technologies effectively, we must have a clearly defined set of activities and a clear process flow for those activities. This is not AI.
AI is vastly different to this prescriptive process in that it does not require a clearly defined workflow. Instead, AI is focused on learning and reasoning, the way humans do, in order to help support a decision or outcome. AI does not work on a rigid and routine assembly line with next-to-no variability, but instead operates by learning functional approximations to solve problems based on previous instances of a similar feature set.
In securities finance, we still see significant straight-through processing and efficiency gains to be had using automation and then leading onto AI. This could be through the use of trained predictive models for detecting violations related to the Central Securities Depositories Regulation, or front-office fraud detection through activity analysis with historical data made available through automation techniques.
Myth #2: AI and automation will replace jobs
Forrester estimated that 10 percent of jobs were to be automated in 2019. It is to be seen if this happened. Gallup reports that nearly half of all jobs in the US are at risk of automation. It is true that commercially motivated tasks are becoming more automated than ever before. However, it is rare to find an employment position outside of a factory that involves completing one menial task on a repetitive basis. In this way, AI and automation should be complimentary technologies to today’s work force and not a displacement. For example, we often encounter tasks that centre around binary segmentation of data; should this entity receive credit? Should this entity be filled for a sought-after special stock? is this entity a risk? In order to make these decisions, although it often seems like all of this data should already be joined up, a human must collect hundreds of data points from systems relevant to the decision, organise the data (often mentally), and determine the relevant course of action. A combination of automation and AI can gather all the data, perform a functional analysis and recommend a decision with a certain level of confidence. The informed human with tacit knowledge of a particular industry is often still required to make the final decision. This relationship not only serves the client, but also helps to train the AI further.
In this way, the AI augments the decision-making process of the human by completing and performing a regression over accurate, historical data. Humans are still at the conceptual helm, still using emotional reasoning, client relationships and implicit knowledge while adding value to a business where it is truly required and valued.
AI, as it currently stands, cannot contextualise or empathise. AI cannot retort ‘that must be hard’ to empathise with an emotional client who may need to cover a position in order to fulfil an important commitment to their counterparty. It cannot determine that a transaction was fraudulent or that a new counterparty is unlikely to settle on time, based on a set of features it hasn’t been trained on. It cannot decide the most socially acceptable time to make a work-related phone call. This is predominantly why job security is not endangered as we enter the new epoch of workplace automation and intelligence. Augmenting an employment position with AI will serve to reduce the impact of human irrationality in decision making while redirecting this emotional intelligence into other, more beneficial areas – such as liaising more effectively with clients and stakeholders.
How to successfully implement AI
These commonly held misunderstandings demonstrate the challenge for organisations to leverage and excel with AI. Many employees lack a deep understanding of what AI means for them and their firms, and this uncertainty can breed fear, especially when it is uncertainty regarding job security. Any new technology leads to a potential sense of loss, because there is always some task that the technology will address that will be ‘taken’ from an employee.
This is especially true in the case of AI, because it seems to be appropriating everything. However, as demonstrated above, this is merely a myth. That AI is ‘appropriating’ is a dangerous ideological viewpoint to adopt, instead, ‘augmenting’ is more accurate. The pertinent question then - is how do firms successfully adopt AI while augmenting the valuable work of employees and simultaneously reassuring them? Here are three ways to start:
Take on the myths: Leaders and stakeholders need to explore what they seek to gain from the integration of AI into their business. They need to consider to what extent they want or are even talking about AI or solely automation? In short, don’t beat around the conceptual bush. Deal in facts and not conspiracy – be transparent with employees about the tasks that are intended to be automated but explain the opportunity for them to make new, more valuable contributions.
Engage your team. In each use of AI, there will be a point where the machine stops, and the employee picks up the work. If AI is about the relationship between humans and technology, begin by talking to people about the relationship between humans and technology.
Engage your workforce about where the machine ends, and the human role begins. Have your team define those tasks where they could use support collecting and parsing data that will equip them with the right insight to make critical decisions. Traditionally the workforce could not build exceptional client service and relationships at the same time as being high tech, efficient and effective. AI can enable your team to be both high-tech and high-touch.
Acknowledge the loss. As discussed, with technological advancement comes tension. Work to reduce the entrenchment of misguided beliefs through encouraging a mindset shift amongst employees. In the case of AI and automation, this shift is difficult because we are asking employees to do something we have not asked since the birth of the knowledge worker - to be more human.
We want employees to be creative, to have time to explore interesting and profitable avenues of thought while we leave the mundane to the machines. There will still be a sense of loss because we are creatures of habitual nature. But your team, once given this push in the right direction, are more likely to embrace the change and create positive outcomes.
Leaders must engage their teams by articulating their vision of what they see over the horizon. Then we are truly ‘talking to tech’.
Over the past century, technology has vastly distorted and disrupted the landscape of the workplace. From Henry Ford’s assembly line of the 1910s to Toyota’s ‘just-in-time manufacturing’ from the 1970s and, more recently, online help bots for sales and customer support as firms have strived to reduce human-based procedural overhead to add value in more important facets of business. Perhaps the most notorious disruption to the workplace in the past decade has been the introduction of achievable and effective automation and artificial intelligence (AI).
Automation and AI are coming (if they haven’t already arrived) to the financial securities market and will vastly change the way teams function. We are already beginning to see these technologies being integrated into the modern work environment, and this mass propagation has introduced a provocative tension between employees, stakeholders and the technology itself. The automation and AI market is set to reach $70 billion in 2020, according to Forbes, and it is therefore paramount that firms address the origin of the tension felt by their employees which is borne predominantly from a misguided fear of human redundancy.
It is no secret, then, that the cleanest process for integrating automation and AI into critical aspects of business is acknowledging that it is not the technology at the heart of successful AI, but rather the people. By engaging your teams on how you intend to leverage the power of Automation and AI to lift the shackles of cognitive fatigue and stress associated with menial, repetitive tasks, you create an atmosphere of optimism over adversity. Employees are less likely to make disastrous mistakes in data entry and rationalisation and can instead begin to focus on adding value to the business in other, more creative ways.
Due to the manual nature of many processes, the securities finance industry has not been alone in encountering challenges relating to human input errors. Firms often encounter erroneous bookings made by trading desks and persistent idiosyncrasies surrounding static data which take a significant amount of time to understand, let alone, remedy. These can often result in detrimental outcomes for market participants and even the smallest error can leave businesses dangerously exposed down-stream.
Using tools from the modern deep learning and automation arsenal, we can introduce trust-worthy, data-driven solutions to these problems that will engender a symbiotic relationship between AI, automation and employees. This will boast huge advantages such as boosting productivity, creativity and morale. Because AI predominantly excels in pattern matching, it seems that the securities finance industry could be an optimal ‘playground’ for these algorithms. We are confident that in the near future solutions will begin to provide ‘what if’ scenarios and simulations with narrowing latent space to provide collateral management efficacy, among other post-trade processing benefits.
The IBM advertisement strikes at the heart of what needs to be acknowledged – namely that we have an inescapable and intrinsic relationship with technology in our modern lives. This intrinsic relationship aptly justifies the bold use of personification. Technology, like a ‘dear friend’ whom we may indeed write a letter to, helps us complete goals, organise ourselves and connect with the world on the international playing field. In recent years, technology has garnered good press for its ability to connect the world, allow social mobility and simplify our lives. However, current discourse surrounding automation and AI does not reverberate with the same levels of optimism. Instead, a mention of AI or automation is shrouded in numerous layers of mystery and fear of almost mythical proportions. Why is our potential relationship with AI so different to that of other technological advancements in the workplace?
Myth #1: AI is just automation with extra steps
It is extremely important to educate employees on the nature of AI in order to help direct thought processes about the new technology.
Perhaps the most important distinction to make between AI and automation is that of routinisation. Automation is the adoption of an approach that removes routine work from the process flow of organisations. The integration of a factory-like, blue-collar style to white-collar work has created aspects of our jobs that are repetitive and can be streamlined through automation. In order to deploy automation technologies effectively, we must have a clearly defined set of activities and a clear process flow for those activities. This is not AI.
AI is vastly different to this prescriptive process in that it does not require a clearly defined workflow. Instead, AI is focused on learning and reasoning, the way humans do, in order to help support a decision or outcome. AI does not work on a rigid and routine assembly line with next-to-no variability, but instead operates by learning functional approximations to solve problems based on previous instances of a similar feature set.
In securities finance, we still see significant straight-through processing and efficiency gains to be had using automation and then leading onto AI. This could be through the use of trained predictive models for detecting violations related to the Central Securities Depositories Regulation, or front-office fraud detection through activity analysis with historical data made available through automation techniques.
Myth #2: AI and automation will replace jobs
Forrester estimated that 10 percent of jobs were to be automated in 2019. It is to be seen if this happened. Gallup reports that nearly half of all jobs in the US are at risk of automation. It is true that commercially motivated tasks are becoming more automated than ever before. However, it is rare to find an employment position outside of a factory that involves completing one menial task on a repetitive basis. In this way, AI and automation should be complimentary technologies to today’s work force and not a displacement. For example, we often encounter tasks that centre around binary segmentation of data; should this entity receive credit? Should this entity be filled for a sought-after special stock? is this entity a risk? In order to make these decisions, although it often seems like all of this data should already be joined up, a human must collect hundreds of data points from systems relevant to the decision, organise the data (often mentally), and determine the relevant course of action. A combination of automation and AI can gather all the data, perform a functional analysis and recommend a decision with a certain level of confidence. The informed human with tacit knowledge of a particular industry is often still required to make the final decision. This relationship not only serves the client, but also helps to train the AI further.
In this way, the AI augments the decision-making process of the human by completing and performing a regression over accurate, historical data. Humans are still at the conceptual helm, still using emotional reasoning, client relationships and implicit knowledge while adding value to a business where it is truly required and valued.
AI, as it currently stands, cannot contextualise or empathise. AI cannot retort ‘that must be hard’ to empathise with an emotional client who may need to cover a position in order to fulfil an important commitment to their counterparty. It cannot determine that a transaction was fraudulent or that a new counterparty is unlikely to settle on time, based on a set of features it hasn’t been trained on. It cannot decide the most socially acceptable time to make a work-related phone call. This is predominantly why job security is not endangered as we enter the new epoch of workplace automation and intelligence. Augmenting an employment position with AI will serve to reduce the impact of human irrationality in decision making while redirecting this emotional intelligence into other, more beneficial areas – such as liaising more effectively with clients and stakeholders.
How to successfully implement AI
These commonly held misunderstandings demonstrate the challenge for organisations to leverage and excel with AI. Many employees lack a deep understanding of what AI means for them and their firms, and this uncertainty can breed fear, especially when it is uncertainty regarding job security. Any new technology leads to a potential sense of loss, because there is always some task that the technology will address that will be ‘taken’ from an employee.
This is especially true in the case of AI, because it seems to be appropriating everything. However, as demonstrated above, this is merely a myth. That AI is ‘appropriating’ is a dangerous ideological viewpoint to adopt, instead, ‘augmenting’ is more accurate. The pertinent question then - is how do firms successfully adopt AI while augmenting the valuable work of employees and simultaneously reassuring them? Here are three ways to start:
Take on the myths: Leaders and stakeholders need to explore what they seek to gain from the integration of AI into their business. They need to consider to what extent they want or are even talking about AI or solely automation? In short, don’t beat around the conceptual bush. Deal in facts and not conspiracy – be transparent with employees about the tasks that are intended to be automated but explain the opportunity for them to make new, more valuable contributions.
Engage your team. In each use of AI, there will be a point where the machine stops, and the employee picks up the work. If AI is about the relationship between humans and technology, begin by talking to people about the relationship between humans and technology.
Engage your workforce about where the machine ends, and the human role begins. Have your team define those tasks where they could use support collecting and parsing data that will equip them with the right insight to make critical decisions. Traditionally the workforce could not build exceptional client service and relationships at the same time as being high tech, efficient and effective. AI can enable your team to be both high-tech and high-touch.
Acknowledge the loss. As discussed, with technological advancement comes tension. Work to reduce the entrenchment of misguided beliefs through encouraging a mindset shift amongst employees. In the case of AI and automation, this shift is difficult because we are asking employees to do something we have not asked since the birth of the knowledge worker - to be more human.
We want employees to be creative, to have time to explore interesting and profitable avenues of thought while we leave the mundane to the machines. There will still be a sense of loss because we are creatures of habitual nature. But your team, once given this push in the right direction, are more likely to embrace the change and create positive outcomes.
Leaders must engage their teams by articulating their vision of what they see over the horizon. Then we are truly ‘talking to tech’.
NO FEE, NO RISK
100% ON RETURNS If you invest in only one securities finance news source this year, make sure it is your free subscription to Securities Finance Times
100% ON RETURNS If you invest in only one securities finance news source this year, make sure it is your free subscription to Securities Finance Times
