3 basic steps to start-up your AI Business Strategy

Rapid advances in machine vision and language processing are pushing AI into the mainstream. Organizations that strike the right partnerships between people and machines can radically enhance their competitive advantage through new offerings, sharper value propositions, and more efficient processes.

But unlocking the real value of AI requires you to integrate AI into your business and customer strategy deeply so it can be unleashed to advance priority goals and drive growth.

Also, even as organizations rush to implement AI strategies to align with their digital transformation journeys, it’s essential to shift the conversation around AI from “everyone needs it” and start focusing on how to measure the success of these initiatives indeed.

More and more I consider AI as a strategic competency that works better when in tandem with human talent to solve complex challenges, break into new markets, diversifies revenue streams and accelerates operational efficiency. 

For many organizations, AI is already a public face of their business, handling everything from initial interactions via chat, voice and more, but to scale AI is essential to seamlessly combine AI business strategy with a bottom-up AI planning framework, as well as infuse AI into the strategy team itself. 

During the last years I've seen an explosion of interest in machine learning technology and potential applications and I consider a good starting point when applying AI in Business the assessment of Machine Learning technologies for products and business or as a potential investment. 

Machine Learning is meant to be used in the context of a given task, a problem with inputs and a way to objectively assess how right or wrong a defined output is. I used to say that you may not fully understand the technology being used, but it’s fundamental that you understand clearly the task you are assigning to it.

To help you to start, I selected here some very basic first-steps to you consider when defining your strategy for your next AI business initiative.

1) Understand what classification, regression, and ranking is.

Predictive modeling is about the problem of learning a mapping function from inputs to outputs called function approximation.

Classification is predicting a discrete class label output for example, and regression is the predicting of a continuous quantity output given a specific input.

For example, image recognition is a classification task where we have an input image and want to predict the primary subject matter of the image (a photo of a dog, car, etc.).

Regression is about predicting a real numerical value or values from an input, such as predicting the future value of a home or a stock portfolio. 

Ranking is about predicting an ordering of items which is “best” in a given setting; for example, in search ranking, we want to order results that are most relevant for a given query and user profile and history.

2) Understanding your evaluation metrics

After building a machine learning model, if we need to measure the performance of our model, we use some metrics to evaluate them.

Once you understand clearly the task, it’s important to know how your Machine Learning system is being evaluated on that task. Typically, people will define a system evaluation metric that gives a quantitative measure of how well the system does on the task. 

As an example, an image recognition you can report what percent of the time you predict the right category for an image. The Machine Learning tasks all have standard evaluation metrics with which it would be worth familiarizing yourself.

Evaluating your machine learning algorithm is an essential part of any project. Your model may give you satisfying results when assessed using a metric say accuracy_score but may give poor results when evaluated against other metrics such as logarithmic_loss or any other such metric. 

Most of the times we use classification accuracy to measure the performance of our model. However, it is not enough to truly judge our model. In this post, we will cover different types of evaluation metrics available.

3) Apply the right metrics for your AI projects.

Measuring the business success of AI is a complicated question. For some companies, these solutions are labor and cost savings measure and should be analyzed accordingly. Other implementations are tied to revenue generation, while some might not fit so nicely into “hard” metrics.

I've seen many times business who developed very sophisticated algorithms and technology for problems solving, but not developed an accurate evaluation metric. 

Not having a parameter is a terrible starting point. There’s no other way to know if your “super deep learning” actually yields any tangible benefits to your business. When it comes to building AI solutions, or any technology, for business value, I always recommend working with focus and drive by metrics.

It's good to consider that more complex AI technology does not necessarily mean improvements in evaluation metrics; especially in environments with limited data, simple techniques frequently outperform more complex ones. When developing business AI solutions and Proofs-of-Concept, always develop and try simpler methods first.

What's next?

Understanding how developments on your AI task will affect which business metrics and by how much is tricky but crucial because there’s a direct relationship on it. 

If executed thoughtfully, incorporating AI into your business model can benefit your customers in ways that facilitate a better customer experience, better relationships, and create opportunities for new revenue streams.

AI promises to change the way we run our businesses, empowering us to make better decisions more quickly, improving how we as humans run our businesses today, to allow us to be more productive in our work than we otherwise would be on our own.

Are you ready?