Rob believes that its customers and clients should be running 100s of AI experiments in order to deliver pay off on some of them. The belief that there could be one "pinnacle project" that might help a business get to the next level is dangerous.
There have to be 100s of experiments on AI with the assumption that some of them will work out. It would be naïve to believe that all of will work and create favorable results.
The motivation for most companies to get into AI is linked with operational cost reduction or modernization of BI and data warehousing. This is where most of the clients are today. Insight driven as a part of self-service analytics is on the verge of creating some good results. And this is where the AI strategic priority for IBM clients is right now.
The next level is transformation that will lead to new business models. 85 percent of customers view AI as a strategic opportunity.
Some of the example that Rob shared includes the Royal Bank of Scotland achieving a 40 percent call deflection rate with virtual agent. Red Electrica de Espana uses AI to predict power demand for renewable energy, Harley Davidson uses AI to predict and target first time buyers in the US, Autoglass visually categorizes damage and used AI to instantly issue quotes, to name just a few.
There are many examples but the key point is that AI is great with large sets of data where you need to make sense of it. This is where AI can massively help and there will be many new use case scenarios happening across all industries. AI is much more than just training and inferencing, it is about getting value from your data and essentially save or even earn more based on it.
So if you a are business, better bet on multiple research than just one big AI project, Rob concluded.