Open search

    Value-generating AI and data strategy

    The majority of industrial companies have yet to start utilizing artificial intelligence ─ how to get out of the starting blocks?

    I claim that data is the least utilized resource of industrial companies. Few companies in the industry have a strategic plan for utilizing data in developing their own operations, improving the customer experience, let alone commercializing data. Without a strategy, there is also no clear ownership or responsibility for maintaining and developing data quality.

    Data-driven leadership should stem from your strategy

    Substantial amounts of money and effort have been spent on building basic systems to enable operations. Sensors have also been installed. The data generated by the systems is used to a minimal extent; sometimes, it is not even preserved (with or without sugar) in the data reserves to wait for better times. This brings a senior advisor to tears.

    Technology is evolving, and applying it to business problems is becoming more affordable, which is an excellent thing. However, there are a few things to be aware of. Data generated by sensors, MES, automation, and ERP systems gushes from all sources. The sirens that lead the cloud platform trade whisper in our ears the message: "Take all your data to the cloud, turn it into new knowledge capital and make better decisions based on it. It costs only a little. Try and be surprised".

    This approach rarely results in anything other than satisfied cloud merchants and bewildered customers. Personally, I see the matter a little differently. If there is no vision to leverage the data that supports the company's strategic goals, activities could easily remain dabbling with detached experiments. If the pursuit is not aimed at some greater goal, the threshold may become too high for applying the culture of experimentation that has been much on display recently. Traditionally, failure has not been looked at admiringly in Finland.

    Data-driven leadership must always be based on the company's strategic goal and the desired change. Finally, the review arrives at the point where an inventory of available data is made. Sometimes the company has the necessary data available (but not always readily accessible), and sometimes it is discovered that the required data cannot be found, so it must be collected, for example, through additional sensors. Sometimes data also needs to be purchased. Finding data-based use cases and business case ideas and making an inventory of data resources may not be easy. Diverse skills and experience are needed. With this, Insta's experienced advisors can help.


    Measures that change the future

    I've always liked the phrase, "Think BIG, start FOCUSED." However, with the certainty gained from experience, I would add one more descriptive section to it: "Think BIG, start FOCUSED. Plan to OPERATIONALIZE". This last addition is particularly well suited to the area of artificial intelligence. The potential for random AI experiments to end up as profitable production solutions is small. Experiments must have a strong link to the company's strategy and, thus, also clear ownership.

    In addition, experiments need to be designed as if they were going into production ─ even if there is no certainty that this will happen when the trial starts. The explanatory power of the available data will tell you this during the experiment. Planning for production includes elements such as cloud services, data integration, and data quality assurance. No experiment, even a good one, is of any use if it does not go into production.

    "Success is only possible if you build on a strategy that takes into account business, operations, and technology. Measurable goals and indicators need to be set for an AI project. Even individual experiments should always be designed to grow the overall capacity of the organization to utilize digitalization."