The most important thing to consider when building new technology is to listen to the market, build for market needs, and don’t invest time in building features the market is not interested in. If you get to market as soon as possible, you can get feedback from users/market. Technology solutions should be people-centric: build for teams and users, not just for the sake of technology.
I believe the most successful Industrial IoT and machine learning companies of the future will be those that solve specific, repeating problems with a combination of domain knowledge and technology, via services models and excellent, stable multi-disciplinary teams to enable sustainable value for the customer.