Recently I had the privilege to attend The Spark award event (co-sponsored by McKinsey and Handelsblatt) in Berlin. Winners of The Spark award are digital startups whose ideas have the potential to cause ground-breaking changes within industry.
It was a great event, full of exciting companies with new approaches. What was striking to me was the fact that there was literally no startup which did not pitch on the basis of its AI-based business model, its unique algorithms or the ability to make machine learning the core of its functions. In most cases, the startups did not even know what the results of the algos would be – would they be profitable? Feasible?
This may be a stimulating exercise for a startup, but it is not a valuable approach for established companies. Algorithms are often a “black box” process for a very specific business problem, and do have challenges to scale vertically and horizontally.
So how can you use an innovative approach like these startups did, but for an established business? How do you take the ideas and the approach, the unique functions coming from AI and start to seed them into your business function and products?
Here are five suggestions that will help:
- AI only works on the basis of data. If you have the data, you can innovate with AI – if you don’t have it, get it! How? The Internet of Things is the most important enabler for new acquiring data. Why? It is THE way to understand how your products are used in the field. This opens the door to a complete new universe of detailed real-time data you not only need to digest but also use to lay the foundation for new ways to capitalize on the data.
- Pay attention to the quality of the data. There is no magic tool that will fix bad data; AI will not solve it for you in a magical mystery way. The same applies to AI; bad data will create bad analytical models. The data needs cleansing and preparing to ensure it is good enough to use in models and algos. This is especially true in an established environment with lots of history.
- Business transformation will be driven by the ability to act in real time. Relevant data processed by AI will give you the fundamental tools to act but to make a real difference you need to have precise and timely decisions based on the analysis of your data. The value of the data erodes over time, so you must change your processes to react to the outcome in real time. This will require complete digital process chains.
- Don’t misuse your data scientists; they have special skills which you should develop. Let them use their insights and advice to become a new source of innovation for your enterprise. You won’t need hundreds of them but you will need them – so please don’t treat them as data monkeys. The data scientist should work hand-in-hand with business experts because there is no good data-driven insight without domain know-how.
- Take an evolutionary approach to AI – open up your data sources to supercharge existing products or create new ones. You can only really differentiate by applying intelligence to the data; information needs to be augmented by AI in order to evolve.
Supercharge your strategy using AI and dramatically improve your products over time. There is no AI “magic” required; just a well-planned step-by-step approach.