According to a Cisco survey, most companies aren’t quite ready yet to work with artificial intelligence. The problems start at the base: 81% of respondents admit that their data is stuck in different silos. With data scattered across the organization and a lack of decent data management, companies quickly hit a wall when taking their first steps in the world of AI.

It’s no surprise that companies don’t have central data management. In most organizations, every department has its own database and its own way of gathering data. Each team only adds the information that is relevant to their job. This system creates multiple silos with plenty of data, but none of it good enough for artificial intelligence. AI needs a clean, complete and central database to live up to its full potential. 

How do you fix data silos? 

Most software companies have APIs that make a connection between multiple data systems possible. This can help to gather data from different sources and bring it together in a central workspace. Departments can exchange information, even when they use different systems. 

Don’t leave it all up to technology to fix your scattered data, though. Your employees will need to help too. For one, they shouldn’t just add the information relevant to their team. They should also enter additional data, important for other departments in the company. People also need to pay attention to how they add this information: it should become a habit to enter data in the same way, to avoid duplicates and to ensure everyone can find the right information. 

Can I still start an AI project, even if my data isn’t up to par? 

Gathering and cleaning data, and setting up a decent central data management system takes time, effort and money. So does this mean you can’t experiment with artificial intelligence until that tedious job is done? Not at all. The key is to start small and experiment. 

You don’t need the entire company’s data to set up an AI project. Does your sales department have a decent set of data? Then start from there. Prediction models, for example, can be useful for your sales team. Using historical data, AI can forecast future sales. AI can also help with customer segmentation, dividing your customers into groups with similar characteristics. 

Of course, a small data set limits AI. The technology won’t deliver in-depth insights. But it does give you an idea of the potential and the opportunities. If you can get this much out of a small data set,  imagine what AI is capable of when you give it all the data. It’s certainly worth the effort of cleaning and centralizing your company’s data.