Artificial intelligence (AI) at your organization: if AI won’t improve your organization should you deploy it? This is more of a rhetorical question since many examples of AI exist—from the old rules-based expert systems to more modern deep learning applications—providing significant positive impact.
However, just as you would not typically get into your car to drive to your neighbor’s place next door, you might not want to deploy AI in cases where it makes little sense. The key is to determine how to know when leveraging AI makes most sense. As with almost any decision, it is about optimizing the cost/benefit ratio.
For the decision of driving your car next door, you can quickly calculate the costs and benefits by considering questions such as:
- How fast do I need to get there? If the neighbor lives 100 feet away, there’s not much benefit. However, in a rural area where you neighbor lives a mile away, driving may have its benefits.
- What kind of effort am I trying to reduce? When walking 100 feet carrying a pie, then the level of effort is only minimally reduced by driving if at all. However, if you agreed to bring your grill, then maybe using transportation is the better option.
It’s all about understanding what benefits you expect at what costs. To get there, you need to understand your status now and what you want it to be in the future.
Leveraging AI for Your Enterprise Strategy
In a recent HFS Research blog, the author discusses the problem that most organizations don’t rightly understand that AI is not an “app that can be installed and rolled out” like an enterprise application for sales force automation or Enterprise Resource Planning (ERP). The reality is that AI has use cases where it provides benefits and cases where the benefit is not as strong.
Applying AI in general to a given problem should go beyond using the technology that is in vogue.
Just like we say here at Parascript, we fit the form of AI to solve a task in the most appropriate manner: applying AI in general to a given problem should go beyond using the technology that is in vogue. Most AI, as Everest Group points out, is better at solving “input tasks” where the ability to crunch a lot of data to identify the hidden patterns and provide meaningful output is the goal.
AI-fueled Effort Reduction
AI won’t solve every problem in the most effective, reliable way. For some problems, the best option is to use a basic form of AI such as a rules-based expert system approach. While other problems have too many variables to manually encode and therefore, might make use of machine learning. Some problems do not have a good-enough input data set for machine learning forms of AI to effectively build a reliable model so a rules-based approach may be more effective, at least in the short term.
Ultimately, as HFS Research points out, there are a lot of steps to undertake in order to arrive at the most appropriate use of AI in an organization. AI is neither an off-the-shelf product, nor a panacea, but can be effectively reduce the level of effort required to complete specific tasks with high quality results.
If you found this article interesting, you might find this eBook useful, Machine Learning for Advanced Capture: