An AI implementation can be a technology project. AI-driven transformation cannot. Whether an organization truly realizes the potential of new technologies depends primarily on its business objectives, how it manages change, the skills of its employees, and the principles that guide AI development.
In large organizations, conversations about AI are increasingly moving beyond the question of whether new tools are worth exploring. The focus is shifting to where AI can be applied, how to move from pilots to broader implementation, and how to ensure employees actually use it.
Paweł Rybak, whose work focuses on digital transformation, points to another important shift: “C-level executives are generally already convinced.”
The numbers reflect this trend. According to the Stanford AI Index 2026, in 2025, 88% of surveyed organizations reported using AI in at least one business function, up from 78% a year earlier. At the same time, 79% reported using generative AI in at least one business function.
However, using AI does not automatically mean that a company can translate it into better business performance. There is still a significant amount of work between deploying a tool and changing how an organization operates.
Paweł Rybak highlights the progress already being made in this area:
“We’re seeing more and more proof-of-concept projects actually move into implementation.”
The next challenge, then, is moving beyond individual implementations toward solutions that work in day-to-day operations and deliver measurable results.
Start with the “why,” then choose the technology
It’s easy to start with the tool. A company buys licenses, rolls out Copilot, or builds its own assistant. More teams begin testing solutions, and the list of initiatives gradually grows.
The problem arises when it becomes difficult to define what specific outcome these efforts are supposed to deliver.
Paweł Rybak puts it bluntly: “There’s this YOLO approach: let’s implement AI. But what are we actually trying to achieve with it?”
That question should be answered before choosing the technology. The objective doesn’t have to be complicated. It could be reducing customer service response times, minimizing manual data entry, preparing proposals faster, or reducing errors in a specific process.
Two things matter: identifying the area you want to improve and defining how you will know whether the change has delivered the intended result.
Instead of a broad goal such as “we’ll implement AI in sales,” a company could set a specific target: reduce the time needed to prepare the first version of a proposal from two days to a few hours. A clearly defined objective makes it easier to determine whether AI is actually needed, what data and integrations will be required, and how to measure the outcome. It also helps determine whether the solution is worth developing further once the pilot is complete.
This matters because the number of experiments alone says little about their impact on the business. McKinsey’s 2025 research found that most organizations were still in the experimentation or pilot stage, while only some were translating AI adoption into enterprise-wide results.
That’s why, before starting a project, companies should define not only what they can build, but also what outcome would justify continued investment in the solution.
Get access to premium content
Leave your email address to unlock the full knowledge base, get notifications about new episodes, and co-create Euvic Talks with us.
We respect your time and inbox. We only send practical knowledge and invitations to valuable events. You can unsubscribe at any time.
