The term Artificial Intelligence (AI) is one that elicits confusion, simply because of the broadness of what it can cover, particularly in the context of data science. AI has many subcategories and levels of complexity that are very different in what they do from a business point of view.
Automation is a well-known subcategory of AI software that follows pre-programmed rules to run processes. However, this is a different level to AI software that displays self-learning capability through machine learning.
AI can mean different things to different people. Different industries might give you different descriptions and explanations of what it is, depending on the possible use cases. That’s why it’s impossible to provide a definitive business description of what AI is.
Freeing up time with AI
Dr Ali Rezaei Yazdi understands this well. He is the Post-Doc Data Scientist at Rimilia, a company which develops intelligent financial software that harnesses artificial intelligence and machine learning algorithms to automate accounts receivable processes including cash allocation, credit management and bank reconciliation.
To Dr Yazdi, AI in the near term is about giving people the time and freedom to do what they’re good at – using our conscience and cognitive capabilities to think, consider, analyze, make informed decisions and provide wisdom.
“Machines through automation can perform the boring and time-consuming tasks that can waste a lot of your time,” he says. “Computers through AI can take the information you submit to provide you the ideal place to eat in a second. Why spend time searching, when with a program you can be provided with a suggestion?
“Understanding that you can use machines for these boring, repetitive tasks was very important to me, and was one of the reasons I got into AI. Life is very short, and we’re living in a world where we’re faced with lots of options, possibilities, data and devices.
“Using a restaurant example – going to all the places you can go to could take hours, days if not months. Why not save time through software using information you’ve submitted based on yourself, your family, or your friends?”
Three levels of Robotic Process Automation
Under the broad definition of the term AI comes Robotic Process Automation (RPA), which is already being used to automate back office functions in many work environments. There are three levels of RPA that business leaders should understand:
1. Simple process replacement
This is considered the entry level of robotics, which tends to be process automation or the replacement of manual repetitive task by a software generated program. We’re already seeing this in finance departments as part of accounting software, and it has been in maturity for the last 10-15 years.
Not only does software replace data from documents or add to a workflow, it can learn and complete transactions. It could for example read fields in document and know what fields to populate, or use an algorithm based on previous known transactions to make software-based decisions that would take a human a long time to do.
This is where software automation gets to a level associated with more advanced forms of AI, as it can take in different and complex data sets, using algorithms to not only learn but make decisions much faster than humans.
Dr Yazdi says, “For large businesses, RPA can increase productivity, saves costs, increase efficiency and make processes simpler and easier.”
He also explains why RPA is particularly valuable to businesses that want to make more money through innovation and creativity, through the creation of new products and services:
“Businesses can free more time to innovate by using machines to take on repetitive time-consuming tasks. Humans can spend more time on creative tasks which need reasoning, cognition, thinking and wisdom.”
Artificial intelligence in 2019: A handbook for business leaders
Get the must-have guide for senior leaders looking for practical guidance on harnessing the power of AI in your organization.
Why digital transformation is key to getting started with AI
Dr Yazdi believes that the key for business leaders when it comes to making use of AI effectively and efficiently, is to make sure they understand their business goals and objectives and where AI comes into it.
“There are businesses of all sizes that aren’t fully clear about what their goals and objectives are,” he says. “What are they trying to achieve? Have a clear strategy and goals at the end of the year you’re looking to meet, and consider what AI can do to help. A clear action plan, considering the constraints and limitations your business might face.
“It could be that your business needs a simple automation of processes from looking at your planning. In other industries, you may be looking at advanced AI using concepts such as machine learning or the mimicking of human behavior. It all depends on your business goals – bring your information together and see what level of AI is needed to meet them.”
To power AI capabilities to maximum efficiency and effectiveness, businesses need to undergo digital transformation, which allows them to discover insights that matter most, embedding them into the software customers and employees use to engage, and continuously learning and measuring through results.
Forrester Research says this means a need for businesses to invest in ‘systems of insight’ with cross-functional insight teams, repeatable insight-to-execution processes and a digital insights architecture.
“Successful digital transformation is at the heart of artificial intelligence – centralizing, organizing and structuring data,” says Dr Yazdi. “AI is based on information – you need data processing and analytics to provide the real intelligence. Digitization is part of that – taking naked data and turning it into clean on-demand data. To convert data into knowledge and wisdom, you need to upgrade your technology.”
Employ data scientists in business that actually understand finance
Dr Yazdi has a PhD in Artificial Intelligence, and is currently leading a three-year project run between Rimilia and Aston University to research AI and big data trends in payment behaviors and trends using AI, machine learning and other mathematical and computational modelling tools. His job is to extract valuable insight from data analysis to help financial teams make informed and fast decisions, using techniques such as probability analysis, decision trees, neural networks, regression models and other classification models, as well as advanced data visualization tools.
He has an interesting way to describe what he does, and why data science is crucial when it comes to delivering insight: “In healthcare, if you want a cure, you go to a doctor. If you want to achieve a goal and get valuable information from data, you need data science.”
He also has advice for businesses employing or looking to employ data scientists into the finance team or wider business. “Data scientists need to understand that business is very different to university life. They need to have a clear idea about the context of the work they’re doing in their individual industry.
“In finance, a good data scientist or AI developer won’t necessarily need the knowledge of a credit controller or accountant, but they must have a very clear idea of what a credit controller or accountant will be thinking about.
“Only if they understand what’s going to happen in the mind of a financial executive, can they produce a result, knowledge of visualization that will benefit the CFO or CEO. For example, if you’re in a position at a company like Rimilia, you need to understand the context and process of the finance back office.”