A new AWS study shows that companies worldwide are prioritizing investments in generative AI, creating new executive roles, and adopting hybrid development strategies—while security concerns are increasingly taking a back seat.
According to the "Generative AI Adoption Index" from Amazon Web Services (AWS), 45 percent of surveyed IT decision-makers plan to make generative AI their top priority in 2025, ahead of traditional security tools, which only 30 percent favor. The study surveyed 3,739 IT leaders across nine countries, including Germany, the US, and Japan.
Mid-sized and large companies are especially eager to invest in AI, while smaller firms continue to put security first. For most organizations, easy integration into existing workflows is the key factor in choosing tools—especially in highly regulated industries (65 percent). In sectors like financial services and education, advanced features (56 percent) and data privacy (48 percent) are also top priorities, much more so than in less regulated fields, where only 34 percent consider security features relevant.
Chief AI Officer becomes a key executive role
As budgets shift, a new C-suite position is emerging: 60 percent of companies already have a Chief AI Officer (CAIO), and another 26 percent plan to add one by 2026. CAIOs are responsible for coordinating AI strategies across departments. Companies like Amazon and Airbnb have even brought AI experts onto their boards.
However, structured change management strategies are lagging behind: only 14 percent of organizations have such a strategy in place, although that figure is expected to rise to 76 percent by 2026. Even then, a quarter of companies will still operate without a formal transformation plan.
Moving from pilot to full integration
Ninety percent of companies are already experimenting with generative AI. But only 44 percent have moved beyond the pilot stage and are working toward full integration. On average, companies ran about 45 experiments in 2024, with only 20 expected to be deployed in production by 2025.
The biggest roadblocks: a shortage of skilled workers (55 percent), high development costs (48 percent), and issues like bias and hallucinations in models (40 percent). Access to clean, high-quality data is seen as the key to reliable AI applications.
Companies focus on training and aggressive recruiting
Fifty-six percent of organizations have already rolled out training initiatives, and another 19 percent will follow by the end of the year. Still, 52 percent say they don’t actually know what training their workforce needs. Implementation issues (47 percent) and budget constraints (41 percent) are also common challenges.
At the same time, 92 percent of companies plan to actively recruit AI talent in 2025. In 26 percent of cases, at least half of all new hires will be expected to have skills in generative AI. Requirements are especially high in the ICT (35 percent) and manufacturing sectors (28 percent).
Out-of-the-box solutions trump in-house development
Only 25 percent of companies plan to build their generative AI capabilities entirely in-house. Most are opting for a hybrid approach: 58 percent want to develop their own applications using pre-trained models, 55 percent prefer fine-tuned models, and 40 percent rely on ready-made AI solutions.
This openness to standard solutions is especially high in traditionally data-sensitive industries like finance (44 percent), education (45 percent), and ICT (43 percent)—mainly because of the advantages in cost, speed, and functionality.
Third-party providers play a central role: 65 percent of organizations will partner with external vendors for AI projects in 2025—15 percent entirely, and 50 percent through a hybrid model that combines outside partners with internal teams.
