The AI Trap!

I have seen many business leaders hoping for AI to magically change their organization's face. That's not going to happen. The sooner they learn this the better it is for them and the business. AI, with all its pristine and glory, need several things in place to deliver the impact at organizational level. And it happens over a period of time (and not in a jiffy that the leaders expect it to be).

This article is not to undermine the power of AI (and related Machine Learning technologies), but to urge its use with discretion. Not every situation warrants AI application. Also, not every organization is AI ready. With full understanding of the field- right skills, data maturity & organization culture, AI will bring power to your organization.

Rise of the Machines

It was only recently that Artificial Intelligence seemed like an intangible, futuristic concept. It has been decades in making- the past 90 years have seen both renaissances and winters for the field of study.

Thanks to recent advancements in data, computational prowess and new age algorithms, today AI is defining the singularity moment. Not only has AI infiltrated into our personal lives with self-driving cars and personal assistants. It is changing the game in enterprise as well by way of adaptive marketing and cyber security, to name a few. Its application to business problem solving is growing in leaps and bounds. 

AI technologies- namely Machine Learning and its subset Deep Learning are pushing the AI envelop. The AI field is practically exploding now. If entrepreneurs and business innovators are to be believed- AI will be in just about every product and service we buy and consume.

This exponential growth of technology has led every organization to scout for AI

It is very much required for organizations to keep tab on latest developments in the field of AI and adapt its systems where possible. Yes, there are situations and use cases where AI can help organizations. BUT not many organizations have the right context, knowledge and data to handle AI. Moreover, not every situation warrants AI. As the hype around AI technologies grow, more and more companies and executives fall prey to the AI Trap.

The AI Trap

People often throw the AI terms around in an attempt to win buzzword bingo. Many service providers in this space are wrapping their traditional solutions with AI cover (with little or no change in the solution). Within organizations, data scientists use complex AI algorithms to impress the audience or because everyone wants AI. No business wants to be left behind. I recently had a client asking to ‘do AI’ because they wanted the sexy stuff. But neither the use case warranted AI, nor was the data sufficient to train AI models on.

It's important to pick the right algorithm and right data for the job at hand.

You know you are in The AI trap when you use it where you don’t need it. Within organization, this happens largely for two reasons. a) you try showcasing your AI prowess or b) you obsess over model performance over business value.

You must realize, finding the model that fits your data best- a problem in predictive and prescriptive analytics- sometimes requires AI and related machine learning, but doesn’t always. Some more traditional statistical analysis works better in many cases. For example- a bi-variate correlation is best suited to find the behavior between variables of interest; predictive modeling using logistic regression is best suited for choice modeling in certain scenarios. These (traditional) analysis are adept at not only finding what variables impact the outcome significantly, but are also able to decipher to what degree as well. That is, these methods are not a black-box, a requisite that most of the business problem warrants.

How does one know if they are in AI Trap                       

Check for following things

1.    Parsimony: Is there a simpler solution than AI

2.    ROI: Does AI justify the investment for this project

3.    Data: Do you have the required data (quality & quantity) to justify for AI

4.    Organizational Readiness: Is your organization ready for AI- the culture to conduct business AI way

Use the right tool for the job, and don’t feel ashamed if you’re not using Artificial Intelligence. Don’t let the current AI hype make you use the wrong tool for the job. Understand the applications of AI and where they are most useful.

Remember, the role of analytics is about adding value to business and not just using the fanciest toys. Choose the right weapon for the battle- AI or otherwise. May the best player win!