Reusable AI: Raise the treasure
Millions of Europe’s SMEs in Covid-19 economy can not wait years until future AI centers of excellence have graduated and trained millions of missing data experts. Let’s look at Fast AI cooking to move DataOps teams at lightning speed.
Artificial Intelligence can be like cooking
Europe has 23M SMEs, and according to the EU’s digitization index, more than 1.6M of them plan or already started their digital journey to become data centric.
Terms such as “data centric” ship with so many different meanings that we decided to share our definition to be clear from the very beginning.
We focus on the functional phases how data is used to drive business actions (e.g. decisions or automated commands to control machinery): track, collect, aggregate, analyze, optimize and action. Companies who achieved to integrate all these phases into their business processes is said to be “data centric”.
Key to reach the action phase is analytics. What do we mean? The full spectrum of machine intelligence from business rules and SQL queries to deep and machine learning to time series and text processing.
Many survey studies, however, reveal that most SMEs cannot adopt this kind of machine intelligence by their own: Lack of skills, limited budget, and missing data strategies, to name a few common barriers.
Suppose there is a need for 1.6M AI solutions and every solution requires a team of 5 experts (which is very optimistic) working for a period of 3 months (which is also optimistic).
We would need at least 2M affordable experts to respond to these market needs within one year. Should we create many AI centers of excellence as often suggested, and wait until this approach provides any business impact for smaller companies?
Let’s compare two manufacturers. A small one and a large enterprise: Both companies may have the capability to deploy sensors and operate an IoT platform. This covers the three phases track, collect and aggregate of the data process and both companies can monitor what happened with their machinery.
When it comes to e.g. forecasting energy consumption or retrieving indicators for production downtimes, the large enterprise is able to adjust resources and mitigate risks beforehand.
The small company is not. And it is obvious that this firm cannot wait years until the amount of graduated and trained data experts is available to optimize its energy load and reduce production risks.
AI is like cooking
We need a completely different strategy to successfully transform millions of Europe’s SMEs into data centric businesses. And, as a prerequisite, we must find the right balance between AI perception and reality:
It is a common misconception, that AI solutions are expensive and time-consuming, require research and scientific work and always solve unique business problems.
No doubt, there are some business cases, where this perception is true. However, in most cases, business problems are similar, and we can build AI solutions like cooking: with pre-built AI ingredients and along reusable AI recipes.
Regional AI solution providers are key to reach millions of SMEs. They can be considered as AI restaurants: Small DataOps teams move at lightning speed and perform many AI projects in significantly shorter periods of time for more and more business users.
In the next article, we describe why this metaphor is no technical illusion and how the treasure trove of reusable AI can be raised and used to boost AI project rates.
Originally published at https://www.linkedin.com.