Making AI software available is easy. Actually gaining productivity from it is much more difficult. According to Koen Van Beneden, Managing Director of HP Nordics and Benelux, companies therefore need more than a collection of licenses. A comprehensive AI strategy rests on three pillars: personal productivity, the use of corporate data, and AI in their own products and services.
Almost every company is experimenting with generative AI today. This often starts with access to a general chatbot or an assistant such as Microsoft Copilot. Employees can use it to write texts, summarize meetings, process information, or prepare presentations. This can save individual employees time, but a broad rollout does not automatically lead to measurable productivity gains. Without clear applications, guidance, and training, there is a risk that AI will primarily become an additional cost.
According to Van Beneden, companies realize they cannot escape the technology. “Every customer we speak with understands the potential of AI and how disruptive it can be, both positively and negatively. They know that if they don’t get started with it, someone else will.”
First pillar: personal productivity
The most visible application of generative AI is supporting individual employees. AI assistants can help with drafting emails, analyzing documents, improving texts, and collecting or structuring information. Van Beneden therefore sees applications such as Copilot and other productivity suites as the first pillar of an AI strategy. Local AI tools can also play a role in this. These perform certain tasks directly on the laptop, meaning information does not always have to go to the cloud and no tokens are consumed.
However, technology alone is not enough. Employees must learn which tasks AI is useful for, how to give good instructions, and how to check the outcome. Otherwise, companies risk paying for licenses that are only used to a limited extent or incorrectly. HP says it has currently given about half of its own employees access to various AI tools. By the end of the year, the entire organization should follow. The emphasis is not exclusively on software.
“Everyone receives mandatory AI training and is encouraged to try and test things,” says Van Beneden. “We realize that users will work in a very different way when they have those tools. We don’t want half of our people to suddenly be unable to keep up.”
Second pillar: making corporate data usable
The second pillar goes beyond the productivity of a single employee. Companies have vast amounts of internal information, from manuals and customer files to sales data and technical documentation. Generative AI can make that data more accessible and reveal connections that are otherwise difficult to find.
But that is precisely where organizations often run into their existing IT problems. Information is scattered across different systems, data is not consistently structured, and it is not always clear who should be allowed access to which documents. “Many companies are coming to the painful realization today that the basics must be in order first,” says Van Beneden. “You have to structure your data well and ensure that the company’s tooling is up to scratch.”
Placing an AI model on top of poor or incomplete data does not solve those problems. It can even magnify them, as employees are more likely to be presented with incorrect or outdated information. Access rights also remain essential. An AI assistant must not show an employee information that they would not be allowed to consult through regular systems. Only when that foundation is in place can organizations build applications that truly utilize their own knowledge. According to Van Beneden, HP is conducting more than two hundred experiments worldwide involving AI and corporate data. These projects investigate how the company can use information more efficiently and organize processes differently.
Third pillar: AI in products and services
Finally, AI can also become part of what a company offers itself. For HP, this includes researching how AI can improve the performance of printers, headsets, and other devices. Companies that are not technology producers can also undertake this exercise. An insurer can use AI to process files faster, a retailer to help customers more personally, and a manufacturing company to detect defects earlier. This shifts AI from an internal efficiency tool to a part of the business model. This requires closer cooperation between IT, product development, and the rest of the organization. The question is then no longer just how employees can work faster, but also how AI can change the service or the product itself.
Many companies are coming to the painful realization today that the basics must be in order first.
Koen Van Beneden, Managing Director of HP Nordics and Benelux
Technology, data, and people
The three pillars are interconnected, as employees need suitable applications and training, and corporate data must be usable, up-to-date, and securely accessible. At the same time, organizations must investigate where AI enables new products, services, or processes. A standalone Copilot license can be a starting point, but it does not constitute a strategy in itself. Without training, usage remains superficial. Without a solid data foundation, corporate applications remain unreliable. And without concrete objectives, it is difficult to determine whether the investment is truly paying off.
Van Beneden compares the current AI wave to the introduction of the internet. Back then, companies also had to invest in connections, Wi-Fi modules, security, and new ways of working. Those costs ultimately were no reason to ignore the technology. “You could have decided twenty or thirty years ago not to participate in the internet,” he says. “But you probably would have been left quite alone. The same applies to AI today.”
The challenge is therefore no longer deciding whether a company will use AI. It is about implementing the technology purposefully, bringing employees along, and turning experiments into applications that demonstrably create value. Making AI software available is easy. Actually gaining productivity from it is much more difficult. According to Koen Van Beneden, Managing Director of HP Nordics and Benelux, companies therefore need more than a collection of licenses. A comprehensive AI strategy rests on three pillars: personal productivity, the use of corporate data, and AI in their own products and services.