In this article, I would like to share the simple decision-making tool named AI Canvas and being used in MBA graduates at the university of Toronto’s Rotman School of Management, by professor Ajay Agrawal, University of Toronto.

Before going into further details, it is vital to declare that AI / Data Science / Prediction is neither a quick process to be complete by days nor a once-off exercise.  It is a continuous process and expected to take up to 1 year for getting good result.

There are 7 elements to be used to work out AI with AI Canvas and they are:

1. Prediction – have our own Crystal Ball
2. Input – consolidate data from different source with data cleansing
3. Judgment – apply Cost & Benefit Analysis
4. Training – create an accurate model that answers our questions correctly most of the time
5. Action – perform decision implementation
6. Outcome – maintain KPI or archive the goal
7. Feedback – improve your AI by outcome data as another data source

Diagram 1: The AI Canvas – 7 elements

After we have the basic concepts about the elements for the AI Canvas, it is better to put the elements into a logical flow with the sequence as the diagram (diagram 2)

Diagram 2: Flow of the AI Canvas elements

First of all, the starting point is to collect the data.  The data is expected to be cleaned and with high quality.  Then, the data will be feed into the training model for the prediction.  Once the training result is good & tested, the insights will be transformed into action with the risk control by Judgment.  The judgment could be by human or another machine driven engine.  Next, it is expected to see the outcome after the action.  With the outcome in-place, it is expected to have feedback as a data source for better prediction in the future.  By this cycle, it is expected to have better and better results with more data and better training.

The above approach is suitable to work with the Total Quality Management (TQM) approach for any organization.  Both AI and TQM should be a continuous work to improve the organization everyday.

TQM is a system of management based on the principle that every member of staff must be committed to maintaining high standards of work in every aspect of a company’s operations.

Diagram 3: Total Quality Management (Plan-Do-Check-Act)

However, we may not apply AI into everything but AI is helping us to solve a problem better & better.  It is always important to educate the whole organization with the AI Canvas (just like a toolbox) before moving ahead to AI.  I do believe the AI Canvas is not just fit for applying AI but also the digital transformation for an organization.  In a nutshell, a business could survive only if continuous growth and improvement with the competitors worldwide by the Internet.

Samuel Sum

Data Science Evangelist (CDS, SDi)

Vice President (AS)