Friday, March 30, 2018

Impact of AI in business

Recent article in the Economist claims that across many industries, many large companies are deploying AI (machine and deep learning) at scale, and achieving cost savings of around 5 to 20% in different business processes from customer service to supply chain.

Deploying AI on any scale is not trivial. While deep learning is not hard to start, it takes adequate investment to train deep networks both in terms of computing power as well as skilled resources to execute.

In any case it is more likely that some businesses will start specializing and commercializing these offerings in specific verticals and develop offerings to sell across players. I suspect in the long term, as  these companies gain more customers, it will cease to be a competitive advantage.  On the contrary, it will reduce barriers to entry for new players to enter into the segment. If that happens, expect that many more industries will get commoditized.

With these trends, expect firms "selling spades to gold diggers" to do well. This will include companies developing new GPUs as well as companies that have a lot of data to run algorithms on. These could be folks such as Thomson Reuters .

Here are the trends that could take root.

1. GPU architectures for deep learning will keep improving

2. Performance per unit power will keep going up.

3. Companies will specialize in specific AI algorithms and productize these for the rest of companies in an industry

4. New players will enter markets lured by cheaper operating models that prove to be unsustainable in the end.

5. Data aggregators who sell data to AI companies will start coming up.

It will be an interesting space to watch.

No comments: