Are AI Agents the Way to Go when Improving the Company Performance?

Listen as an audioblog (05:42).

Assuring the improvement effectivenss is key when aiming at a higher process improvement levels.

Many companies already implement, or at least consider, the use of AI agents to improve the performance of specific processes. Some of the companies have even provided guidance at various management levels that improvements should be based on AI based solutions, and a certain objective such as a 30% productivity increase should be achieved.

From a process change perspective, the above is justified as any change to the process will do ("green flag"). But from a process improvement perspective it is already questionable as real improvements may not show up despite all the bells and whistles ("yellow flag"). From a high-performance process improvement perspective the above is a solid no-no ("red flag").

Why is that so?

If you only have a hammer, everything will look like nails. Having different kinds of hammers (AI agents), won't take your company that far, as a set of different hammers will still only fix the nails. What if the main problem or improvement object is not hammering the nails at all, but doing something else?

Firstly, the essence of high-performance process improvement is to assure the improvement effectiveness which means that the company improves the right issues. Secondly, once the improvement effectiveness is assured, the improvement efficiency should be assured which means that the issues are improved in the right way. These key improvement disciplines should be done in a classy way, swiftly and cost-effectively. In this blog let's focus on the first improvement discipline.

An AI agent is in the end only a piece of technology. And to put technology, whatever it may be, in the correct perspective, it is only one component that affects the process. Not the only one. After all, a process is the interaction of people, technology, information and materials to produce a desired output (product or service). How well this interaction works in real-life affects directly the process performance (time, quality and costs). The combined sum of the performance of all processes and activities in the company is the company's total performance.

To add up complexity, a company and therefore also the related processes need to satisfy the core of the company's stakeholders, i.e. the internal or external customers, the personnel and owners. If these are not sufficiently satisfied it is hard to satisfy other stakeholders such as investors, society, environment, etc.

Improvement effectiveness means in practice that all relevant issues affecting a certain process are disclosed and prioritized in terms of implementability given the real-life improvement potential and organizational constraints. These constraints include stakeholder needs and demands, the strategy, time, financial considerations, change resistance and required efforts to improve the company culture, knowledge and skills. To make things more difficult, top management, middle management and employees or workers have different considerations and perspectives that should also be considered and harmonized properly.

To arrive at the correct conclusion what should actually be done and how require always the use of tacit knowledge, which means by definition that this kind of knowledge is not in a digital format that can be accessed via a database. Only the best improvement objects with the best improvement traction and real-life outcome should be selected for implementation. This input is unfortunately missing when implementing AI agents using the "technology first" approach, which means that a lasting positive effect is basically based on luck or a "gut feeling". In the long run, it is not a good idea to build the success of a company on luck or a gut feeling, as you may run out of good luck any time, and your gut is not always on top of things.

In conclusion, it all boils down to what ambition level you want to run your improvement efforts on. If you intend to just change the process — and want to gamble — then AI agents and a technology first approach may suit your company's needs. If you want to run the improvement efforts on a higher ambition level, then such an approach is not advisable. The scarce resources should always be deployed to verified improvement objects with the best real-life improvement traction for your company's specific process. These verified improvement objects could, of course, contain AI agents as a solution. Thus, the first thing an enlighted and forward looking company should do is to understand the essence and concept of improvement effectiveness, and make use of high-class process improvement plans that comply with the requirements of high-performance process improvement. The VISTALIZER Report process improvement plan provides a good solution in this regard. After all, let the others gamble — whereas you build sustainable success systematically. With or without AI agents.