11. March 2025

Software Diversity and Artificial Intelligence: How AI Can Solve the Fragmentation of Work Processes

By Jana Blankenhagen

Software diversity: mastering the juggle

Eine Frau mit lockigem Haar steht vor einem kastanienbraunen Hintergrund und trägt eine Jeansjacke über einem weißen Hemd. Sie lächelt leicht und hat die Arme verschränkt.

Jana Blankenhagen

Chief Human Resources Manager

A scenario all too familiar: It's Monday morning and, as usual, you're sitting at your laptop with your first coffee of the week, trying to get an overview of what's on the agenda for today and the next few days before the first workday really gets going.

As always, the day begins with a quick check of the unread e-mails in the mail program. Then you open Microsoft Teams, for example, to take a look at upcoming meetings and get an overview of all projects and tasks in the countless Planner boards. Next, you log in to the specialist applications to check your other tasks and follow-ups, e.g., in the document management system or enterprise resource planning tool, and check whether all workflows are running correctly or whether there is an urgent to-do. The last thing to do is to quickly check who of your colleagues will be there and what appointments they have, who you can reach when and how between your own appointments to clarify the most important things—or simply to ask whether everything is OK or how the last weekend was.

Before the workday has even properly started, it already feels like you're juggling a chaotic mix of balls—different in size, feel, and color. Each ball has its own character, much like the diversity of people around you. In the world of everyday office life, transferred from sports to technology, a term like software diversity describes it well for me.

What is software diversity?

In my interpretation, software diversity describes the multitude of different software solutions that are used simultaneously in a company. These solutions include software systems in the following categories and examples:

  • E-mail and communication: Microsoft Outlook and Teams, Gmail, Slack, Zoom
  • Document management and process management: ECM systems such as enaio® and yuuvis®
  • Specific specialist software: CRM systems such as Salesforce, ERP systems such as SAP, industry-specific work tools
  • Personnel management and time recording: Workday, SAP HR, Personio

In practice, this means that office workers often switch between many different systems to complete tasks, manage information or communicate.

Software diversity leads to fragmentation of information and working time

This variety of software solutions leads to a major problem: the fragmentation of information and processes across several tools, which in turn leads to increased fragmentation of working time. Employees lose valuable time by constantly switching between different tools. Important information is spread across various programs, which makes it difficult to maintain an overview and reduces efficiency. There is also a risk that tasks or deadlines are overlooked because they are "hidden" in one of the numerous systems.

There are various studies and reports that look at the number of software systems and tools that the average office worker uses on a daily basis. According to a report by Industry Dive in May 2023, office workers use an average of around eleven different software tools per day to get their work done.

Another report in the Harvard Business Review describes how office workers switch between applications on average about 1,200 times a day. This adds up to around four hours per week that are needed for reorientation after switching between applications.

The associated fragmentation of information and processes, as well as the differences in the operation of each individual software, not only require increased attention during use, but also increase stress and workload, because each software is structured differently, looks different and leads to a different user experience overall. Software is inherently diverse—and rightly so, as it must meet a wide range of technical requirements: an e-mail client is different from ERP software. Working with them generates different thoughts and emotions:

"Where can I find the document again?", "When was that e-mail sent? I can’t seem to find it quickly in Outlook.", "If only I didn't have to keep jumping between all these different programs!"

How could the problems of software diversity be solved with the help of AI?

What would it be like if you sat down at your computer in the morning and your personal AI assistant opened up? The virtual assistant immediately shows you a complete overview of all the topics that need attention during the course of the day. They are prepared and arranged in a logical sequence because the AI has already

  • checked all appointments, workflows, resubmissions and deadlines, regardless of which system they are in, and because
  • all documents and information have already been logically organized by the AI and are instantly available when needed—no searching required, and because
  • the AI even knows about an employee or customer request that you had almost forgotten about yourself. This is because the AI created the task automatically and linked all relevant e-mails and notes.

As you work your way through the day, you realize how much more relaxed you are. The constant stress of keeping track of all the different conversations, appointments, tools and information is a thing of the past. Instead of getting bogged down, you can focus on the essentials: your actual work and, above all, the people you work with.

Over time, you also notice that the AI assistant not only organizes the day better, but also learns from your own working habits. If you regularly work on a project, the AI automatically compiles all the necessary programs and documents before you even need them. It also gets to know your preferred working hours and suggests that you complete smaller tasks at times when you are less productive. This way you can get the "really" important things done in the productive hours.

It quickly becomes clear that this automation and individualization requires complex system integration and the right interfaces in all systems. These are often missing, but would enable AI to seamlessly link data and processes from different tools and platforms. More on this later.

Less stress, more productivity: How an integrated AI decodes software diversity

This is precisely where the use of artificial intelligence has great potential, but is also one of the biggest challenges. A well-integrated AI could be able to seamlessly connect all of an employee's software applications, enabling a more efficient workflow. It would consolidate all relevant data and digital documents from various specialist applications, highlight key priorities, and support efficient day planning. It would dynamically adapt the plan when new tasks arise or priorities change. By minimizing redundant and inefficient switching between a multitude of software applications, AI reduces the associated mental strain and improves productivity.

However, the challenge lies in the connectivity of the systems, as different applications often have incompatible interfaces and data formats. Smooth integration therefore requires not only technical adjustments, but also close cooperation between software providers in order to create standardized data interfaces.

How to train AI to support us in dealing with software diversity

AI learns to support us in dealing with the variety of software tools by analyzing usage data and recognizing patterns in our working behavior. The AI is able to:

  1. Through supervised learning and reinforcement learning, track usage habits and determine which software is used, how often, and when, as well as identify the data linked to specific tasks and the relevant information from various tools at any given time.
  2. Understand priorities—based on deadlines, frequency and context in the documents in which the AI is typically trained through "supervised learning" and "natural language processing (NLP) ".
  3. continuously adapt by means of online learning. Artificial intelligence learns from new tasks and interactions in order to provide more precise recommendations and automation.

The ability of artificial intelligence to learn and adapt is an important criterion for its successful use. The AI would have to learn how each individual employee works in order to make meaningful suggestions and plans. This learning process can be complex, as different working styles and priorities need to be taken into account. In addition, work processes and software systems change over time. AI must be able to continuously learn and adapt in order to remain effective in the long term. This requires a high degree of flexibility and continuous development.

Does everything work seamlessly when cross-system AI is integrated into a platform that serves as the foundation for all your software systems?

No! Of course, there will be colleagues who will be skeptical. Some fear losing control over their work. But if you do it right, you quickly realize that the AI is merely supportive. As a human, you always have the final say and can reject recommendations or adjust the assistant's settings. In addition to the natural skepticism towards new things, there are other issues that are challenging but can be mastered with the right approach.

Navigating the challenges

There are a number of technical, organizational and ethical challenges in the development of such an AI that integrates all tools and simplifies software diversity in everyday working life:

1. Integration and interoperability

  • Variety of software: The biggest technical challenge is the integration of the numerous different tools and applications used in a company. These tools often come from different manufacturers and are based on different technologies, protocols and interfaces.
  • API interfaces: Not all software solutions offer open or well-documented APIs, which makes integration more difficult. Even if APIs are available, they must be compatible and stable in order to ensure smooth communication between the systems.
  • Data formats and structures: Different applications store and process data in different ways. The AI should be able to correctly interpret, consolidate and process data from various sources, such as follow-ups, subscriptions, calendar formats, chats, e-mails and documents, without information being lost or misunderstood.

2. Data protection and security

  • Access to sensitive data: To function optimally, the AI would also need to have access to many different systems and data sources, including potentially sensitive business data and confidential personal information. This poses a significant risk to data protection and requires a high level of trust in AI. At the same time, employees may be allowed to see the results of the AI, but not have access to the associated data. This makes it difficult to assess the accuracy of the results. The legal system would have to be developed in such a way that it minimizes this contradiction.
  • Compliance and legal requirements: The AI must ensure that it complies with the applicable regulations, such as the GDPR (General Data Protection Regulation). Access to data should always be limited to what is necessary and must remain transparent and traceable.
  • Security risks: A central AI that accesses many systems could be an attractive target for cyber attacks. A robust security concept must be developed to prevent unauthorized access to data.

3. Scalability and performance

  • Dealing with large amounts of data: Depending on the size of the company and the number of tools used, large volumes of data can be generated. The AI must be able to process this data quickly and efficiently in order to provide fast and near-real-time support.
  • System load: If AI is integrated into multiple tools and platforms, this could place additional strain on the systems. It is important to ensure that AI does not cause a significant slowdown or disruption to workflows.

4. User-friendliness and acceptance

  • Adaptability: Employees often have to adapt to new technologies and working methods. The AI must therefore be designed to be user-friendly so that it is easy to understand and operate. It should integrate seamlessly into existing workflows and not overburden users.
  • Acceptance: Employees may be skeptical about AI monitoring their day-to-day work and making decisions. It is important to build trust by ensuring that the AI acts transparently and comprehensibly and shows that it actually makes work easier.
  • Flexibility: Different departments and employees often have individual requirements for their tools and work processes. AI must be flexible enough to understand these different needs and provide customized support.

5. Costs and resources

  • Development and implementation: The development of such a comprehensive and versatile AI is costly and resource-intensive. Companies must be prepared to invest in the necessary infrastructure and expertise.
  • Maintenance and updates: After implementation, the AI needs to be regularly maintained and supplied with updates in order to stay up to date with the latest technology and integrate new tools or work processes.

6. Ethical and social implications

  • Autonomy vs. control: An AI that manages tasks and priorities for employees could affect their sense of autonomy and self-determination. It is important to ensure that the AI is supportive and that control remains with the employees.
  • Dependence on the AI: If employees rely too heavily on AI, there is a risk that critical thinking skills and awareness of certain processes will be lost. The balance between automation and human decision-making is crucial.

The use of AI pays off: for more productivity and less stress

By reducing the need to constantly switch between different programs, there’s more time for creative and strategic tasks, as well as meaningful collaboration with colleagues. And even those who are initially skeptical will slowly begin to appreciate the benefits of AI. The entire company benefits from a clearer, less fragmented working environment thanks to AI.

At the end of a financial year, when you look back on your own work, you realize that this change means more than just a new technology. It has fundamentally improved the way of working and helped to focus on the essentials again.

Do you have any further questions?