Most manufacturing businesses have large amounts of data at their disposal. In the chemical industry in particular, companies have taken important steps towards using AI and analytics. But do not expect a large army of data scientists. Such an investment is not feasible for any organization and there are simply not enough profiles on the market. Therefore, it is important that companies get their existing engineers and operators on board of the AI train.
In technical terms, this is called ‘democratizing AI’: making the technology widely available and incorporating techniques into day-to-day operational processes so that more people can work with them. Of course, it is still necessary to have data scientists but they will be more effective if they can focus on areas where they can really make a difference.
No more time to play
Artificial intelligence and machine learning are not new technologies. Data analytics has been the expertise of SAS for more than 44 years. In manufacturing, both concepts have become trendy buzzwords that everybody likes to use. Yet few people understand what you can really do with these technologies. As a result, experiments often remain a Proof of Concept (POC). COVID-19 has changed the world we live in and for AI too playtime seems to be over.
Unless a POC is industrialized, it is no more than a research object. After all, predictions are worthless until they lead to actions. Whoever invests in data analytics, wants to see rapid results. This means that companies must be able to quickly industrialize such a POC and then replicate it. This is only possible with an integrated platform on which many people work together.
This used to be the sticking point, but fortunately many managers are now starting to realize that they can apply AI in multiple domains. In the past, they often used separate techniques for different departments, but in doing so they limited their own scalability. By enabling more people to work with the same technology, flexibility increases and a model can also evolve much faster to respond to changes in the market.
In addition, the real value of AI is not in building a model for one installation but comes when you can replicate that model for other installations. Suppose it takes about 100 days to implement the first model in an operational process, then its replication should be in done in no more than 10 to 15 days.
Building bridges
For SAS, democratizing AI has become an important foundation to build on. More than a seller of technology, SAS wants to be a partner for customers within a wider ecosystem. No one has better knowledge of processes than the engineers and operators who work with them every day. Here lies the key to unlocking the true potential of data analytics.
SPG Dry Cooling, for example, is a manufacturer of cooling installations for power plants. Data analytics has allowed this company, with headquarters in Brussels, to extend its services. After selling an installation, they never received feedback from customers in the past. Now they collect the data that enables them to improve customer services, for instance to predict when an installation needs maintenance and even increase the energy production of power plants.
Many companies are still developing maturity but now need to accelerate. That is why SAS builds bridges between data scientists and engineers, while supporting the development of the right competences. If you can get these two profiles to work together, they will discover that they are looking at the same things in a different way. By merging both visions, you create a new dynamic. You can only experience the real power of data when you add context to it.