These workers pretended to use newly introduced algorithmic technology, but they could never expect what would happen next…

Reading Time: 5 minutes

The introduction of new technologies creates new possibilities for ways of working and organizing, but the implementation of new technology is not always without problems. Therefore it is crucial for business leaders to understand what is happening under the surface of organizational change and look beyond the hype of new technology to make the right decisions when implementing digital technology like AI. This study: Make Way for the Algorithms: Symbolic Actions and Change in a Regime of Knowing by Stella Pachidi, Hans Berends, Samer Faraj, Marleen Huysman, gives an interesting look under the surface of this organizational change when implementing AI technology. In this blog, I will discuss the key findings and asked Stella Pachidi to share some interesting takeaways for business leaders.

Silent resistance

The introduction of new technology is not always without problems. In the case of AI’s introduction in the workplace, AI is seen as a tool for producing knowledge and threatens how employees develop and use knowledge in a specific work domain. In such cases, employees can resist its implementation. In the case described in this research article, the authors saw another form of resistance happening, and this resistance led unintendedly to the full implementation of the suggested technological change. So good news! The technology was fully implemented… But in this case, you can ask yourself if the full implementation is desirable. Or was this a case of office politics gone completely wrong? Let me tell you a little more about this case, and then let us look at what was happening under the surface.

About the case: TelCo is a large organization offering telecommunication services to a broad range of customers. On the one hand, the actors in this research are the business-to-business sales department that targeted medium-sized enterprises, and the Customer Intelligence department (technologist) on the other hand. The technology that was introduced is a data analytics sales tool developed to identify sales opportunities based on predictive modeling and optimization algorithms. 

The snowball 

Implementation of this data analytics sales tool not only meant a shift in how the account managers worked, but it was also seen as a threat to their sales practice because it deemed their expertise redundant. But instead of avoiding the system or resisting it, the authors describe that the employees symbolically conformed to the new technology and pretended to use it without actually doing so. This symbolic conformity led unintentionally to the full implementation of the suggested technological change and to the account managers’ redundancy.

How can symbolic actions from the account managers and the technologists actually contribute to making the technology appear effective in the management’s eyes? See this cartoon to simplify the symbolic actions that happened (see the full article for the full description).

Looking under the surface

To explain this surprising outcome, the authors state that technological change is associated with deeper changes in developing and using knowledge and evaluating people, actions, performance, etc. Those aspects are deeply interrelated, and because they are interrelated, introducing a new technology like AI will have substantial consequences. It is not just a matter of increasing efficiency or cutting down costs. There will be deeper transformations.

To be more specific, the authors suggest a regime-of-knowing lens that analyzes serious challenges happening under the surface during technology introduction. See the box below for more details on this theory.

What is a regime of knowing? Three aspects of a regime of knowing can be distinguished: The knowing practices that shape how actors develop and use knowledge in a specific work domain. For example, Account managers generating highly contextualized knowledge based on personal contact with the customer, using intuition and experience. The valuation scheme is used to evaluate performance, actions, people, objects, and ideas. Example: Having a personal and trustworthy relationship with the customer is essential to perform well in sales. Authority arrangements that offer socially sanctioned ways to structure collective activity, organize work, and coordinate tasks to ensure skilled performance. Example: The account manager should control the sales process because s/he is in contact with the customer. This includes interfering with the closing of the deal. 

Ripple effect of ticking the box

Looking back at the cartoon, let us zoom in on what happened in each scene and how the aspects of a regime of knowledge are interrelated and were affected by the account managers and technologist’s symbolic actions.  

[1-2] The account managers symbolically visit the technologist’s kickoff meetings, although they only went here because that was expected. This impacted authority arrangements. From that point, the account managers cannot ignore the data scientists. 

[3] Then, the account managers symbolically conformed by pretending to use the analytics tool, registering sales opportunities to indicate that they used analytics to generate them. They intended to defend their ways of working and producing knowledge (knowing practices). However, this made the analytics tool appear as an effective tool for finding sales opportunities (valuation scheme).

[4] Because the account managers registered sales opportunities with an indication, this led to an increase in registration numbers; the data scientist celebrates this as a model use success. And use this (and by avoiding thorough analysis) as a symbolic way to advocate the analytics tool is a success, legitimizing the new valuation scheme. Analytics appears to be an effective way of working!

Fast forwarding in time (full description of all the actions you can read in the full article). [5] For management, this apparent success of the predictive model confirmed the superiority of the new way of working and led to the layoff of most of the account managers. 

Symbolic dance

The more the account managers pretended to use the analytics tool, the more the data scientists could claim the tool’s use as a success. In turn, this increased the pressure on account managers to increase their symbolic conformity. This dance of symbolic conformity and advocacy, ironically, accelerated this radical change. This accelerated a radical change in the work practices in the sales department. The organization went through a substantial change by deeming all their salespeople redundant and outsourcing the sales process. Ironically, the analytics tool became the main source of customer knowledge for the organization, even though its effectiveness hadn’t really been tested. 

How can you use this knowledge when implementing new technology?

So what takeaways do the authors have for you as a business leader implementing new technology in your organization? One thing that is a pitfall in a lot of cases is that you should not be blinded by the promise of the introduction of new technology. Choosing to implement digital technology is one, the process through which it is implicated is as important. In this example, the process overlooked the impact on the knowledge. I asked Stella Pachidi, first author of the research article for some takeaways.

4 takeaways for business leaders

  1. Best of both worlds: Consider how to incorporate the advantages of the analytics/AI tool together with the useful idiosyncratic knowledge of employees (in this case, the account managers).
  2. Engage the users early: Engage the users of the technology quite early on in the design and implementation stages. This will help ensure that their expertise and knowledge are incorporated in the tool. Have the developers of the technology collocate with the users to better understand where they are coming from and how they produce knowledge.
  3. Understand what is happening: You cannot manage the technology implementation process from your ivory tower. You need to engage with the workers and understand how they interact with the new tool in practice rather than rely on a set of metrics to assess the tool’s success. When you try to manage by metrics, you have to expect that people will engage in symbolic actions and other evasive maneuvers. 
  4. For employees: You cannot simply ignore and pretend that you are okay with the changes implemented in your organization. At some point, your symbolic conformity will come back to bite you. Instead, raise your concerns and try to collaborate with management (and the technologists) to implement a change that will be beneficial to all.

Suggestions for future reading

If you want to learn more about the intended and unintended consequences of AI in organizations, read this article:

The full research article: Stella Pachidi, Hans Berends, Samer Faraj, Marleen Huysman (2020) Make Way for the Algorithms: Symbolic Actions and Change in a Regime of Knowing, Organization Science,

Author of this blog

Christine Brauckmann works at KIN Center for Digital Innovation where she brings science to business in many different ways. One of them is through the KIN business network and an other is by writing blogs about academic research with a focus on business implications.

I would like to thank Stella for her feedback on this blog.