- Experimentation with unfamiliar technology is not encouraged–especially for traditional accountants. Algorithms and AI are inherently probabilistic, and there will be many situations where the machine comes up with a “wrong” answers. But we forget that people also give “wrong” answers. AI decreases the risk of an incorrect conclusion if we are given all the collectible facts.
- Another barrier to entry is the attitude of tax professionals on value creation and clients.
The perception of value was the toughest barrier to overcome and it took me seven years. The adoption of AI and machine learning allows me to focus on the analysis of the data that is collected rather than the low-level tasks which machines can do, AI can accentuate the value that the individual contributes. I also understand that there is a limit on the types of tax problems which merit the investment in developing an AI solution.
Issues arising in the tax advisory field are characterized by a unique fact pattern, which makes the issue as a whole impossible to solve efficiently using AI. But, if it is broken down into its smaller parts, some aspects fall into the repetitive, high-volume, fuzzy matching, classifier problem that is suitable for AI and machine learning.