Artificial intelligence in SMEs: 5 tips for implementation
A thought-out introduction and the appropriate use of AI can bring SMEs real added value.
When implementing AI or AI systems, SMEs should not rush into anything. The opportunities must also be recognized and internalized by employees. After that, you can introduce AI with a clear roadmap and the use of relevant data.
Investment by European startups in artificial intelligence (AI) is rising sharply. That is the conclusion of a recent study by management consultants Roland Berger. According to the survey, investments in the European AI ecosystem grow by up to 50 per cent annually. In 2019 alone, 218 artificial intelligence companies were founded in Germany.
According to another study by the auditing firm PricewaterhouseCoopers GmbH (PwC), the gross domestic product could increase by 11.3 per cent by 2030 due to AI technologies. These figures demonstrate the high relevance of AI for companies.
Taking the horror out of AI.
When companies plan to introduce AI solutions and systems, it changes many employees’ everyday working lives. Uncertainty and fear may appear, which find expression in concerns about one’s job . Employees ask themselves the question, “Will algorithms replace my job?”
Many people associate AI with something threatening. Something mystical that they can’t grasp.
But how do you succeed in taking away the terror that comes over employees when someone mentions AI?
Companies and executives are in demand. They need to show that AI, or AI software, like predictive analytics, is not a threat but can make daily work easier.
AI can relieve the burden of routine tasks, such as leaving more time to take care of customers and pursuing value-creating activities.
SMEs should, therefore, communicate openly and transparently about the introduction of a planned AI system. Playing with open cards, not leaving employees in the dark, pointing out the advantages of AI, and making the personal benefits clear are essential factors crucial to the success of the introduction of AI and its acceptance by employees.
Develop a strategy and roadmap for the use of AI.
At the outset of implementing an AI solution, such as for sales , SMEs should evaluate what benefits the measurement will bring them: What can AI get us, what can it do, and what can’t it do? It may make sense to cooperate with a data scientist.
The next step is to describe what the AI solution should achieve in concrete terms. For example, sales could mean gaining new insights about customers from existing data to increase sales chances. Or: Using existing customer data with an AI solution for predictive analytics, derive the probability of which customers will buy additional offers (cross-selling and up-selling) or which customers may drop out or churn.
HOW CAN ARTIFICIAL INTELLIGENCE WORK FOR ME?/a>
Preparing for the introduction of AI.
In this phase, it is a matter of adequately qualifying the employees to use AI solutions in everyday business for the new requirements and challenges. Structures in the company organization and the distribution of tasks may also have to be adapted.
Experts advise testing AI systems in pilot projects and phases of experimentation and gain experience under real conditions before implementing the board’s applications. After implementation, you should regularly review the use of AI solutions.
Check availability of data and data quality.
Without relevant data, AI systems cannot develop their benefits. The background is that AI-powered processes refer to machine learning.
In this process, a computer program analyzes data and uses self-learning algorithms to try to identify specific patterns and regularities in that data. Machine learning aims to intelligently link data, recognize correlations, draw conclusions, and make predictions.
Therefore, it needs historical data so that AI can learn from it and thus provide the company with long-term competitive advantages. SMEs should find out what data they have and how well it is maintained before introducing AI.
In the end, the decisive factor is knowing how reliable the data is.
Letting AI software “train” (modelling).
Modelling involves “showing” the AI software relevant sample data from which an objectified model for machine learning, a “thinking” template, so to speak, is ultimately generated. For example, AI software can be “shown” data on a specific target customer group’s purchasing behaviour and data on their age and income.
The system tries various algorithms until several “trained” models emerge at the end, from which it chooses the most suitable one.
With this approach, however, you may discover that specific data is insufficient or insufficiently processed. You may also include additional data. Therefore, the “training” of the AI software is also a piece of “trial and error”, which means: trying it out again and again until you find the appropriate machine learning model.
Artificial intelligence in SMEs – Conclusion.
A thought-out introduction and the appropriate use of AI can bring SMEs real added value. But only if the companies refer to a current and concrete problem and define how they can solve it with AI.
It is essential to realize that no universally valid algorithm can answer all questions or solve all company issues.