I continue to be underwhelmed by vendor pitches on their AI efforts to date. For some it’s a repackaging exercise of old technology with new marketing lipstick. For others it’s nothing more than multivariate regression modelling which is not AI. This should not be a surprise for most of us in the industry as the hype cycle always precedes the business value.

For buyers of AI solutions, you need to separate the contenders from the pretenders before you commit to any trials much less purchase orders.

For the marketeers the focus at this stage should be laser focused. Don’t try to be all things to everyone. If you can’t explain your value proposition on AI in 2 minutes you have failed.

For the product managers and field experts, be realistic about acquiring high quality data in large volumes. If you stumble here your business case falls apart quickly. Be aware of access to data pipes that may be off limits. And don’t overlook data privacy issues. Data drives the ML engine so unless you can overcome this hurdle you should steer clear of the use case being proved out.

The few vendors of AI in telecom that stand out and continue to push the boundaries of AI include Ciena Blue Planet, Uhana by VMWare, and Nokia. Their success is driven by identifying the data sets and making sure that enough data is available to drive the machine learning models. Each has a compelling, concise business case that is defendable. Finally, each does not over rotate on their marketing. It is on point and probably most importantly has the customer endorsements to back it up.