Artificial Intelligence is making its way to SMEs

AI-based applications are becoming more commonplace in the SME-sector. How can AI enhance operations and increase sales?

The first thing that comes to mind when discussing AI might be self-driving cars and the movie Terminator, while in reality, almost all of us use AI-based services already daily. Some of these include Netflix and, in a Finnish context, Yle Areena (Finnish broadcasting company’s online streaming service) which suggest new shows based on your earlier views. The same technology is used by Google Maps, too, so that the app might suggest the quickest route according to current traffic conditions.

• At its’ utmost simplicity, AI is an algorithm of an information system, which is used to foretell something, states Senior AI Advisor Tommi Pajala from technology company Reaktor.

Global digital giants are making billions with their AI applications, but an increasing amount of Finnish companies are using the technology to boost their business. A typical use for AI is a company’s webshop, where a consumer’s buying behavior and interests can be monitored. Based on this knowledge, the algorithm calculates what additional products could be offered to the consumer.

• If a client buys screws from a hardware store’s webshop, the app will suggest a matching screwdriver.

As technology develops, prices decrease

The use of AI Is becoming more commonplace in the SME-sector, too, as technology develops and prices decrease. Implementation is easier, as AI-related services and know-how are more commonly available in Finland. Tommi Pajala advises starting by mapping out the daily business of the company. Are there routine tasks that could be dealt with faster by a machine? In accounting firms, AI-based worktime monitoring apps are saving time as they interpret scanned invoices and register them to the right account.

• Along with routine tasks, AI is often better than humans in tasks that entail understanding and analyzing complex and multifaceted issues.

For example, an algorithm that observes weak signals from a major mass of data can help a company in decision making, assessing and managing risks, and noticing problems. Through this, for instance, the development of important markets can be monitored, sales steered towards potential clients, or a singular client’s insolvency anticipated.

AI requires adequate data

In addition to algorithms, AI needs enough computation capacity and adequate data to base calculations on.

• Computing power is easy to obtain through cloud-based services. Data, on the other hand, should be pre-existing or obtained from somewhere.

Pajala notes that ‘teaching’ the AI to do reliable foretelling requires plenty of data and information. If a company website is rarely visited, an algorithm used for making recommendations is unlikely to be useful. If there is data on thousands of purchases, AI is better able to make useful calculations.

Pajala reminds us that there are important data protection matters to take into account when taking on an AI-based application; it is important to understand, for example, what form of consent from the user is required to collect and utilize their data.

Consider these things when planning to take on AI-based applications:

1. Start with a small-scale experiment and choose one process which can be simplified with AI. Do not take on too many processes to be changed at once, as it is very likely to not end well.
2. Does your company have enough good quality data to be used? If not, some can be transferred from a pre-existing information system. You can also start collecting the data now and use it later.
3. What type of AI solution functions best for a specific use?

Original writer: Matti Remes
Translation: Helsinki Region Chamber of Commerce