When organisations begin to take an interest in AI, a fear of falling behind often emerges. Many look at their competitors and assume that everyone else is already further ahead and has started to reap the benefits. This creates pressure to develop a dedicated AI strategy—but that is often starting in the wrong place.
You do not necessarily need a standalone AI strategy. AI is a tool—not a goal in itself. Unless AI is at the very core of your business, it should not be treated as a strategy on its own. What you need is a business strategy where AI can be part of the solution. Otherwise, you risk searching for problems that fit a technology you have already decided to use.
As with any new technology, there are risks associated with implementing AI. That is why it is important to start by understanding how much risk the organisation is willing to take—and to weigh that against the potential gains. By starting small, both investment and risk are kept at a manageable level, while the organisation learns and becomes better equipped to make future decisions.
Before implementing AI, it is essential to take the time to understand what you actually want to achieve—and why. This involves analysing the organisation’s needs, constraints, existing systems, and competitive landscape.
It can be a worthwhile investment to run workshops with the teams who will be using the solution, in order to understand their workflows and challenges. It is important to temporarily take AI out of the equation and focus instead on the problems themselves. Once a solution has been chosen, there is a tendency to identify problems that fit it—and that can lead to poor decisions.
Engage with stakeholders, legal teams, and IT to understand needs, requirements, and constraints. Explore the market and document your insights so you can identify the right place to start. A structured discovery process provides a solid foundation for decision‑making.
Rather than making large strategic investments, it is often better to start small using existing solutions. Identify the smallest possible investment that can solve a concrete problem, and move quickly with a proof of concept.
There is no need to find the perfect solution from the outset. The focus should be on testing, learning, and validating whether you are on the right track. Only once a solution proves effective in practice does it make sense to think about scaling, long‑term costs, and partnerships.
At the same time, it is important to avoid becoming locked into a single vendor. If you start with standard, off‑the‑shelf solutions, make sure you retain the flexibility to build more tailored solutions later on.
Experience shows that it is often best to begin with a ready‑made, managed solution that covers most needs. It is easier to expand functionality over time than to fix issues related to compliance or data later.
If you have already experimented with standard solutions, the next step may be to build your own solutions to create greater value. If you are still at an early stage, the focus should be on gaining an overview and identifying the right use cases to start with.
Let go of the idea of a large AI strategy. Instead, focus on your business objectives, understand your needs, and start small. By working iteratively and learning along the way, you can reduce risk and create real value—rather than pursuing AI for AI’s sake.
