*This article was released on the 28th of October in Jyllands-Posten. This edition is modified and translated from Danish to English.
Artificial intelligence is currently at the top of the agenda for many Danish companies and organizations. Curiosity and the fear of not getting started quickly enough with artificial intelligence, also known as AI, go hand in hand. Still, it’s important to think carefully and get off to a good start.
“It’s necessary to have a strategy for how you want to use AI in your organization and, most importantly, how to implement and anchor it so that it’s not left up to individual departments or employees to experiment,” says Anders Boje Hertz, Head of AI & Data Platforms at Intellishore, which provides data and analytics solutions to numerous pharmaceutical companies, the financial sector, global retail players, and other commercial enterprises both nationally and internationally.
Anders Boje Hertz experiences a great interest in AI among top management. “Many companies are afraid that if they don’t hurry to jump on the AI wave, they will be overtaken by competitors, but it’s wise to ease off the accelerator. Setting up 20 projects won’t help if the foundation isn’t in place,” he emphasizes.
Anders Boje Hertz doesn’t view AI as a mere gimmick that will fade away; instead, he sees it as a tool that is here to stay and urges caution. He compares AI technology to the wave of digitalization that transformed paper into data. “As deep-tech experts, we have worked with artificial intelligence in various ways for many years. Now that the value of using AI technology has shifted from programming code to chatbots, there are no barriers to who can use it. The danger is that many dive into using AI without considering how and why they want to use this new technology,” he says.
Although AI technology is now widespread, it still requires careful consideration, strategy, and a plan for implementation and monitoring.
“Deciding that the company should use AI is not enough. There’s a need for concrete plans and approaches to make it a reality in the organization. To succeed in introducing this new technology requires governance and standardization, just like any other initiative,” Anders Boje Hertz points out. He observes that many companies, in their enthusiasm for new possibilities, unleash this new technology without an overarching plan.
“If all departments in a large organization start implementing AI without clear direction and control, the company risks losing both customers and money, and the IT department ends up with a huge cleanup job,” Anders Boje Hertz emphasizes. He explains that he has seen several Danish companies where directors have encouraged all employees to use AI in their workflows without considering whether these employees possess the necessary skills for implementation. “A better recipe for success can be found in organizations that reuse the knowledge and experience that already exists in the company’s data and AI department,” he says.
Before the boss presses the start button, the company should consider how AI should align with the organization’s core values. Which areas should be focused on? Where does it make sense to use AI? How do we ensure data quality and maintain security? Finally, the organization should consider the purpose of using this new technology and how many resources to allocate. “The incentive to use AI should not be to replace employees but to make them more efficient,” says Anders Boje Hertz.
Once artificial intelligence is integrated into the system, significant advantages can be gained both internally and externally. AI can take over routine tasks, freeing employees for other tasks. At Anders Boje Hertz’s workplace, Intellishore, AI has been introduced as an option in the HR department, where a chatbot can answer employees’ questions about everything from maternity leave regulations to holiday pay.
“If the technology functions so that users feel they receive comprehensive answers, it’s not a problem; it’s AI responding. When you get answers to your questions, you don’t think further about who the sender is,” he emphasizes.
One unemployment union took it a step further and introduced a new service where members could submit their CVs and receive feedback and suggestions for improvements based on ChatGPT AI technology. Thus, AI has become a source of new and improved services for members. However, it can also go the other way; the organization risks losing customers and revenue if the technology is not managed correctly. “The worst-case scenario is that customers have a bad experience with, for example, a chatbot and turn their backs on the company,” says Anders Boje Hertz.
Once the company has formulated an AI strategy, the product needs to be tested and adjusted during a testing period of three to six weeks. Can the chatbot answer the customers’ questions, or do they ask something unexpected?
“There should be the same control and respect for AI as other technical initiatives. The implemented solutions, both internally and externally, need to be taken seriously. Security, costs, monitoring, quality control, and backup are just some of the known Machine Learning elements we have learned over the last ten years are necessary to ensure the right governance and compliance,” says Anders Boje Hertz.
Before AI is rolled out in the organization, there must be a plan for how the system will be monitored and developed. Finally, employees need to be introduced to this new world. “It’s not enough for the boss to just press the button. The project needs to be introduced to the employees so they can understand and use the new technology,” he emphasizes.
The new AI era with GenAI is still so new that we are still in an experimental phase. But soon, the service will mature and fully develop, making the benefits more evident. In the end, patience is required because, like any new technology, AI will also have to overcome a range of problems initially.