Why is an AI strategy important?

What an AI strategy looks like depends heavily on the size and goals of a company. Large companies often benefit from planning AI broadly and across all areas from the start. Smaller companies often fare better by starting with individual projects. This allows them to gain initial experience without committing many resources and creates a basis for later investments.

Regardless of size, an AI strategy is always an ongoing process. The path to a data-driven company is never fully completed. Therefore, the strategy should be closely linked to specific use cases and regularly reviewed.

An AI strategy is particularly useful when a company:

• plans its first AI applications,

• wants to make processes more efficient,

• wants to automate recurring tasks,

• wants to address or counteract a skills shortage, or

• is already experimenting with AI.

The earlier goals and priorities are set, the easier it is to make later decisions.

AI strategy in Germany and Bavaria

The federal government and states are also pursuing their own AI strategies – with clear priorities and billion-euro investments. They show: Good framework conditions also determine how successfully AI is implemented in practice.

Germany:

The national AI strategy (2018, updated 2020) was supplemented by the AI Action Plan of the BMBF. By 2025, the federal government is providing around 5 billion euros, including over 1.6 billion euros through the action plan. Focus areas: more AI professorships, stronger practical transfer in companies, and "AI made in Germany".

Source: Federal Government AI Strategy

Bavaria:

With the Hightech Agenda, the Free State has invested around 5.5 billion euros since 2019, of which 360 million euros are specifically focused on AI – for research, AI professorships, and transfer to SMEs. Four focus areas: Intelligent Robotics, Data Science, Health, and Mobility.

Source: Hightech Agenda Bavaria

Developing an AI strategy in 5 steps:

• Step 1 – Identify problem

Sit with your team and collect concrete everyday problems: Which task repeats constantly and consumes time? Where do errors frequently occur, for example in data entry, scheduling, or offer creation? Note how many hours per week each task takes – this quickly shows where the biggest lever is.

• Step 2 – Set goal

Formulate a concrete, measurable goal for your problem instead of a general intention. Instead of "We want to become more efficient" say: "We want to reduce processing time for customer inquiries from two days to two hours" or "We want to cut the error rate in invoice verification by half." This way, you can clearly say after the project whether it worked.

• Step 3 – Choose use case

Choose a task that occurs daily or weekly, is clearly defined, and can be implemented in a few weeks – for instance, summarizing emails, pre-sorting applications, or automatically creating proposal templates. Avoid cross-departmental large projects for a start; they take a lot of time and carry a high risk of failure.

• Step 4 – Define responsibilities

Appoint a fixed contact person who will accompany the project from start to finish – for example, a manager from the affected department or a particularly tech-savvy person on the team. This person plans the timeline, gathers feedback from employees, and reports progress to management.

• Step 5 – Evaluate experiences

Schedule a fixed date right from the start after 4 to 8 weeks to take stock: Was the goal set in step 2 achieved? How has the team adapted to the new solution? Document the findings in writing and decide based on this whether to expand, adjust, or discontinue the use case.

You can find this and many other pieces of information on the AI Compass topic "Strategy" in our download portal: