February 08 2019
By: Aman Mann
Artificial intelligence has become a buzz-phrase of sorts in recent years, and while it’s become hard to separate the hype from its practical potential, AI is rooted in a very realistic notion. Put simply, AI is one of the most advanced concepts in intelligent systems and decision-making in years, and the initial wave of academic research and early engagements with consumer electronics have evolved into real, practical methods and technologies that are positioned to change the way humans make decisions, and eliminate the risk of human error altogether.
AI’s spike in popularity and inclusion in hype machines for tech and electronics does not mean it’s a brand-new idea. It’s been an aspiration and goal of computer science pioneers since the early 1950s, and has evolved incredibly in the 21st century. In recent years, AI has integrated itself with other aspects of our lives as the power behind the chatbot revolution, smart home networks and applications, and assistive voice-based technologies like Alexa and Siri.
In the days ahead, AI is positioned to deliver an immense opportunity to accountants, and promises to bring a new standard of efficiency to their own specific set of decisions and tasks. In the grander scheme of accounting advancements, AI could radically change the profession in its entirety.
Faults of human error
While human intuition is its own cognitive wonder, with particularly advanced capabilities in adaptability and flexibility, it has its limits. The human brain is constantly bogged down by its own inconsistencies and biases, with things like availability bias and confirmation bias proving to be costly in industries and decision-making of all kinds.
The advent of machine learning and AI is actually poised to assist human decision-making, rather than replace it entirely, so throw out the fear of unemotional robots taking over the world (at least for now). Machine learning is here to automate and inform the time-consuming and redundant tasks accountants do on a day-to-day basis, which subsequently frees up their time to focus more on lucrative and in-depth analysis.
The opportunity in machine learning
Machine learning is the application of statistical models and algorithms that actually mirror cognitive strengths like pattern recognition and contextual, specific learning. The most powerful capabilities of machine learning exist in its ability to process large datasets, its adaptability in learning from complex and constantly changing patterns, and its unwavering consistency. The fact that artificial intelligence never gets tired, combined with its complete lack of bias and smaller margin for error, makes it a technology that’s infinitely scalable in many industries.
In accounting specifically, ML’s support of decision-making turns into offering accounting professionals impeccably referential data-driven insights and a combination of financial and non-financial analysis. AI also equips accountants with the tools to solve current and contemporary issues they face, such as the delivery of reliable, cleaner and cheaper data, in addition to the aforementioned capability to let accounting professionals allocate their time to problem-solving, advising, strategy developing, and leading, as opposed to the mundane tasks of procuring and organizing data.
The limits of machine learning
Now that we’ve established why ML is so capable and powerful in the proper environments, it’s time to reveal the catch: ML algorithms and processes are only as intuitive as the data they’re using. If the datasets being input into these models are incomplete, insufficient or are riddled with their own biases, the insights that ML models will spit back out will have the same issues. If you’re using ML and other AI processes to glean results that need high degrees of confidence, this is an issue.
In addition, not every task is appropriate for AI just yet. While the potential is endless, the realities of the platform currently allow it to only execute tasks with a degree of repeatability. This allows the platform to recognize patterns, generalize its learnings and apply them accordingly. The outputs of ML algorithms are predictive and suggestive in nature, which means not all tasks can be handled in this framework.
Problem-solving, en masse
The opportunity embedded in the current digital transformation of finance and accounting is allowing people and AI to work in unison and rely on each other to nurture and contribute in the areas where their strengths exist. The mind-numbing, monotonous minutiae of number crunching in accounting is taken care of, and all that’s left is for accounting professionals to tackle the tasks their cognitive engines are geared for.
Machine learning has proven that organizing and parsing data is no longer a task for the human brain. Machines have finally proven they can do it better, and if we let them spot the patterns and provide us with insights on how to utilize them effectively, we’ll be better for it.