Pretty much all AI/ML customer and transactional models have been on pause for the last few months. As we speak, history is being written. Thus new models can’t be created either as these models rely on the fundamental principle that the past predicts future. So statistical models are not being built unless specifically for COVID impact analysis. What does all this mean for AI and ML models in the future, post-COVID?
The reality is nobody knows. These past months have changed our customer’s behavior. The 70-year-old war veteran never thought he would be ordering replacement tires for his truck, online. I never thought I would order my new car online, and the car would be delivered entirely touchless. I could have never imagined that my 6-year-old daughter would have a special request on her birthday — ‘Mumma, can I meet my friend just once. I want to hold her hands for a little bit’.
As we return to a new normal, consumer habits have changed. The business and social realities are different. The restaurants are opening with new spatial arrangements and barriers. Manufacturing business units can’t operate full capacity to keep the social distance.
Will the consumer ever go back to pre-2020 face to face habits, or somethings have changed forever? When would we get such stability that consumers would repeat tomorrow what they did yesterday?
Unless buying behavior, consumer habits, and transactional patterns stabilize, AI and ML will stay on pause.
1. AI for CFD: Intro (part 1)
2. Using Artificial Intelligence to detect COVID-19
3. Real vs Fake Tweet Detection using a BERT Transformer Model in few lines of code
4. Machine Learning System Design
Does that mean data science is on hiatus? No, not really. Today, more than ever, with shrinking margins, the pressure is on to use data, front and center to optimize decisions. Simple ad-hoc analysis (vs. ML/AI models) will be your friend for massive quick optimizations right now and in the near future.
As economies open up, companies and individuals who can use data they have at hand, quickly, with simple business analytics methodologies, to optimize decisions will flourish. The sales and marketing folks can use aggregate and correlation analysis for current funnel optimizations and customer engagement. The product managers will get most of their TAM analysis done using sizing and estimation. And similarly, most business professionals, if equipped with appropriate data literacy skills, will find much value in acting quickly but with confidence using data-based decisions powered by simple analysis.
Rest assured, AI/ML models will return, as soon as yesterday starts looking like tomorrow.