“The future is independent of the past given the present.”
The above translates the Markov assumption:
Markov Chains help build the concept of decision making in a very simplistic way. It’s also one of the most important deterministic processes used in AI.
How could we bring Markov Chains’ simplicity to business?
Connecting the dots between data, AI, and value creation are essential for working on the right problem.
Professors Ajay Agrawal, Joshua Gans, and Avi Goldfarb of the University of Toronto have introduced a simple decision-making tool: the AI Canvas. The Canvas is a Lean Canvas derived from the popular Business Model Canvas.
The AI Canvas is a way to clarify the seven factors for each critical decision throughout a business:
- Prediction: needs to make the decision;
- Judgment: the value of different outcomes and errors;
- Action: what to do;
- Outcome: metrics for success;
- Input: data needed;
- Training: needs to train;
- Feedback: usage of outcomes to improve.
This will enable you to start identifying the opportunities for AI to be:
1) Better – enhancing performance through skills augmentation to decision making. The impact of human judgment augmentation could save more lives, build safer products or use human’s time in a more efficient manner. Particularly useful in the med-tech, manufacturing and transportation sectors.
2) Cheaper – lowering the cost of prediction by using AI to build the information you don’t have.
3) Faster – delivering a piece of information faster it changes the whole chain of operation. Particularly useful for transportation or manufacturing companies.
Your challenge is to identify where the outcome hinges on uncertainty. Using an AI Canvas won’t conclude if you should make your proprietary AI technology or buy it from someone else. The AI Canvas can help with clarity to understand the prediction, judgment and the action of measuring the outcome, and the types of data needed to train, operate, and improve your intelligence system.
Understanding your lifecycle and examining how decisions are made within a particular context (how the workflows are handled no matter if it’s a person’s workflow, statistical model, or AI model) it’s something businesses need to answer. Taking something abstract like data and results and making them banal and familiar to use (check Google Clips’ journey: https://design.google/library/ux-ai/) requires a great deal of effort and understanding.
As AI is an opportunity for executives in every vertical industry sector, becoming AI-fueled is not optional. If you want to read more on the subject, here are two useful resources: “Prediction Machines: The Simple Economics of Artificial Intelligence“ a book by Ajay Agrawal, Joshua Gans, and Avi Goldfarb. Also, check out Andrew Ng’s “AI for Everyone“ course for Business Executives.
What’s next for AI applications?
As we are entering a new age of AI application, we are seeing the development of programs to enable ‘third-wave AI systems’. It’s where the context and environment of operation understood by machines define explainable AI. Such a program is currently developed by DARPA.