Investing in Artificial Intelligence
Intro
While overall adoption of artificial intelligence (AI) remains low among businesses, every senior executive that I discuss the topic with claims that AI isn’t hype, although they confirm that they feel uncertain about where these disciplines may provide the largest rewards. One premise is obvious to many: To-be-launched AI initiatives must generate business value, revenue, and/or cost reductions, at least at the function level, so that CFOs and other executive committees members are on board. A recent survey by McKinsey claims that a small cluster of respondents from various vertical industries already attribute 20 % (or more) in their organisations’ P&L sheet to AI. The topic is undoubtedly worth a thorough evaluation.
Aiming to speak knowledgeably to provide informed recommendations to the reader, I write from experience, also striving to explore scientific breakthroughs and validated use cases that can be shared. I hope this article may serve as a starting point for any business leader that needs to take a leap forward to sustain competitiveness, or that aims to enhance the quality and impact of their work.
In this article I attempt to advise the reader on different ways to make the first step into the discipline, commenting on the organisational areas where AI can have the biggest impact short-term. Before you read on you should not ignore a crucial point: AI’s end goal is to serve and empower humans to do more, to be better, smarter and happier. Corporations operating more efficiently and generating added value in the market can only be a consequence if this is well understood. I highly recommend a proper understanding of terms (Data Science and AI) as a starting point for digesting this article.
Recommendations
For some time now, CDOs and CIOs have started to use different Machine Learning, and other AI capabilities, without paying attention to the highest possible return on investments. Our work consistently shows that the organisational units that most commonly reap larger rewards when AI mechanisms are implemented tend to be the areas where AI can have the most significant impact. Sounds logical but, what does it mean? Well, AI can have the biggest short-term impact where more money is being spent. So, the first piece of advice here is to follow the money. There are studies from McKinsey claiming that both supply chain management (manufacturing) and sales generation (advertising and marketing on b2c strategies) are the two functional units where AI has traditionally proved to have the biggest impact. Both of these business areas need heavy capital expenditure so it seems reasonable that both areas quickly reap big profits.
AI can have the biggest short-term impact where more money is being spent. So, the first piece of advice here is to follow the money.
The second way leaders can make up their mind about where to apply AI is to simply look at the functions where traditional analytics are already operating usefully and where they may have room to evolve. Simply because AI, through Neural Networks, for instance, may strengthen use cases, by providing additional insight and accuracy, better than established analytical techniques. If you lack the computational power and/or the knowledge on how to apply state-of-the-art machine learning or deep learning algorithms you only need to reach out to data scientists. It may not be valid in some cases, i.e if the additional accuracy is not worth the investment or when a bullet-proof scientific equation already works well and does not require any Machine Learning algorithm. Either way, my second piece of advice: each traditional analytical method currently being used is worth a thorough evaluation since an additional implementation could make a qualitative and quantitative difference.
A third option, looking into potential Robotic Process Automation. You can think of RPA as a software robot that mimics human actions that can handle high-volume, repetitive tasks at scale. It is an evolution of the traditional Robotic Desktop Automation (RDA) that has helped tremendously in the past by simplifying, automating and integrating technologies and processes on employees’ desktops.
The main difference between RDA and RPA is its scope. RDA is implemented in each user device, only interacting with applications and software of that specific user. RPA encompasses multiple users, departments and applications. If you look at the previous graph, you will also be able to see that RPA is associated with doing, whereas AI is concerned with the simulation of human intelligence by machines. RPA is suitable for automating grunt, repetitive, rule-based tasks where humans only spend time actioning them, but can’t improve much over time. For this third option, based on RPA being highly process-driven, I recommend conducting an initial process discovery workshop that may serve as a prerequisite to mapping out the existing “as is” workflows in order to identify gaps and inefficiencies.
Many of our clients believe RPA is a smart safe bet as a first step on the AI stairway and they have cited reasons such as wanting to achieve quick-wins and capturing those low hanging fruits. Our time-to-market for RPA projects at Bedrock is usually a matter of weeks with reasonable costs and challenges, but guaranteeing a measurable return on investment i.e. We have led projects of the kind where human labour went down almost 89%. This is obviously a significant cost reduction, but it also reduces human interaction thereby reducing room for manual errors.
The next step on the stairway would take your business to full intelligent automation, on which I will build upon my fourth piece of advice. AI can help you to automate decision making because modern computational power can outperform humans’ ability to process data and it is your duty to assess if this can help you and other business leaders that you work with. AI applied to power better decision-making has already been called Augmented Intelligence. It is an alternative conceptualisation of AI that focuses on being an assistive role as a cognitive technology, designed to enhance human intelligence rather than replacing it. This means that it could assist not only C-level execs or boards of directors but also any managerial layers in effective decision making. You must plan for building a hybrid collaborative approach where human intelligence and AI work together.
Gartner actually estimated that by 2030, decision “augmentation” will surpass all other types of AI initiatives to account for 44% of the global AI-derived business value. They go on forecasting that in 2021, in the corporate world it will generate roughly $3 trillion dollars of business value. Then, it is your choice if your earnings before interest and taxes (EBIT) falls within this forecasted amount.
You must plan for building a hybrid collaborative approach where human intelligence and AI work together.
My last piece of advice, the fifth, and probably the most important, is to care about your people. Embedding AI across an organisation means a big cultural shock. There has been some paranoia about AI taking over jobs, but it needs to be understood how it is a matter of adapting our society towards advancements. Humans only shape our technologies at the moment of conception, but from that point onward they shape us. This can be seen when we came up with the smartphone because that little device influenced how we communicated with relatives and did business. The same applies to commercial flights or automobiles decades ago. Humans rebuilt our cities and lives based on these breakthroughs and the same will apply to AI.
Many still perceive AI as a job killer when it must be seen as a powerful hybrid (human and robotic) workforce. Getting buy-in for this new “workforce” might be difficult because humans fear the unknown. The correct response from leaders is being open and honest with employees, providing everyone with an understanding of what AI is, how it will be used and what will be the lasting impact on current workers and their lives.
Humans rebuilt our cities and lives based on these breakthroughs and the same will apply to AI.
Moreover, mastering AI requires specialised talent and tech tools, as well as extensive training to ensure proper adoption. The ROI of this last piece of advice is not as tangible and measurable as the previous ones, but it will surely make a difference long-term.
Conclusions
Summing up, I have provided you with objective and useful advice on how to make the first step in the AI journey, sharing five recommendations on how your organisation should make safe moves in the discipline which were:
- Paying attention to areas where big money is being spent, effectively leveraging specific domain expertise.
- Assessing which current analytical methods could be improved. New insights or higher accuracies may quickly surface with advanced ML methods.
- Mapping out repetitive processes that could be subject for efficient RPA. Humans must do intelligent work. Leave repetitive tasks for machines.
- Building a transversal Augmented Intelligence capability. Computers can handle more data than your brain. They also handle it objectively and without getting tired. Make them work for and with you.
- Remember it is all about people: Culture, transparency and robust, efficient processes are the most solid foundation on which to build an AI-powered business upon.
For the past few years, many enterprises have wasted millions of euros on digital transformation initiatives that were not aligned with the real requirements of the business, let alone with individuals’ needs. AI success in 2021 enterprises and beyond will only be possible if you are capable of aligning the right technology, educated employees and intelligent processes to the business’ long-term vision.
If you are leading a company where you are planning to test the waters and make the first step with AI, follow the previous five recommendations and you will not go wrong.
Do not hesitate and move fast!
Start your own ecosystem of AI partnerships and providers. Outsource if you need to do so. Why should you be in such a hurry? Well COVID-19 has exponentially accelerated digitalisation. Companies that are currently benefiting from AI are planning to invest even more in response to the post-pandemic era and inevitably this will soon create a wider divide between AI leaders and the majority of companies still struggling to capitalise on the technology. So the long-term danger for you is not losing jobs to robots, but to remain competitive in your market niche.
