BLOGS
WE SHARE OUR KNOWLEDGE
Challenges in implementing test automation
We conducted a research about implementing test automation within the organizations. Read this blog to learn about the key findings.
The software has changed dramatically from small programs to complex comprehensive IT systems over the last decade. Our lives are increasingly determined by its presence everywhere - right from our toothbrush to our cars. The quality of the software, therefore, stands to be of crucial importance. Organizations, which rely on manual testing methods, have to deal with an increasing pressure on faster time-to-market cycles. Software testers have to work iteratively to ensure the product continues to work even when there are multiple deliveries, resulting in tremendous pressure on the test process. This is where test automation plays an important role.
We conducted research about implementing test automation within the organizations. Based on this research we see that the organizations start enthusiastically with the implementation of test automation, but as time progresses, the enthusiasm decreases and the organization ceases.
Research approach
We examined 17 different projects regarding the implementation of test automation within test and QA organizations of different companies in the Netherlands. As a research method, we used semi-structured interviews. In this case study, the Test Automation Consultants (TACs) who implement or work with test automation were interviewed. This research focused on the question: "What are the positive experiences and challenges with Test Automation within the different organizations in the long run?" Thus, what has happened with the efforts from the past, and what are the changes the test automation will keep running after that one consultant leaves.
Based on interviews, we had positive experiences but faced challenges regarding test automation. The chart below lists the positive experiences and challenges.
Key results of the research:
- There was only one successful application of test automation, which is not a lot and means that test automation needs work, hard work.
- Test automation is started using a project-based approach in various organizations, where the organization and the test professionals are enthusiastic about automating current tests on different levels from a unit to a Graphical User Interface (GUI). The first projects are started successfully, and the agreed deliverables are delivered. The test professionals who are to work with the delivered tool or framework are trained. After the implementations of the tools, some challenges arise, such as maintenance on test scripts or code. This is due to the fact that not all test professionals have the skills to be able to do the maintenance of the test scripts or code. The organizations, however, keep trying test automation and see it as continuous improvement. In an ideal scenario, all the three wheels (as shown in the figure below) need to turn, to have continuous improvement within the organization.
- Most test professionals hired by an organization are outsourced. Thus, test automation remains dependent on individuals or external parties. Transferring the knowledge to the rest of the organization, therefore, becomes a challenge.
- It was also noticeable that not all organizations have a specific test strategy or architecture when implementing test automation. The organizations focus more on automating test scripts instead of thinking of an overall test strategy or architecture for the organization. This results in using tools that do not align with the other tools within the architecture of the organization. And worse - we start testing the wrong things.
The following practical recommendations are based on this research to help the QA & Testing professionals further with test automation.
- Start with the WHAT. What do you want to test and by whom?
- Identify in this the regressions and routine tasks, which can be automated by using test automation tools.
- Maintenance of your test automation is part of your test architecture.
- Make sure that you are doing enough unit testing to avoid finding defects on higher levels.
- Is there enough knowledge regarding test automation within the organization?
- Use it often in your tests to overcome the maintenance bottlenecks.
The learning points from the past are till valid for many organizations.
We know it is not easy to do test automation, however, it is important to have a closer look at your organization and decide what makes sense.
Ethics of Artificial Intelligence
Nowadays, Information Technology is present everywhere. Our lives are increasingly shaped and influenced by it. These developments can be observed across various sectors. Our systems, devices, cars, and equipment operate based on our input, and when issues arise, it is humans who resolve them. But what if our devices could think for themselves and solve problems autonomously? This is where Artificial Intelligence (AI) comes into play.The idea behind AI is to mimic human cognitive abilities. It should enable systems to make independent decisions based on triggers from their environment. The underlying concept of AI is that a system learns to make decisions autonomously by recognising patterns and following algorithms. This means that systems make mistakes and learn from them to become more effective over time. In a way, it resembles an educational or learning mechanism—commonly referred to as Machine Learning.On social media, algorithms are used to make recommendations for videos, advertisements, and other content. While AI can offer significant benefits to humanity, it also poses risks. Because AI systems can make mistakes, these errors may have serious or even fatal consequences in certain situations. Therefore, the legal and ethical frameworks surrounding AI are extremely important. The key dilemma of AI lies in the extent to which we allow our lives to be shaped and influenced by it.
Cybersecurity: From a technical issue to a strategic theme
To Until a few years ago, information security in many organisations was seen as the responsibility of the IT department. Firewalls in place, antivirus installed, and that was it. Those days are now long gone. Digitalisation, cloud solutions, chain collaboration, and hybrid working have made organisations more agile, but also more vulnerable. As a result, cybersecurity has evolved into a strategic theme that directly affects continuity, reputation, and trust.
At Continuous Connect, we see on a daily basis that cyber incidents no longer affect only large multinationals. Medium-sized organisations and (semi-)public institutions in particular are increasingly confronted with phishing, ransomware, or data breaches. Not because they are “careless,” but because their digital dependency has grown rapidly, while the organisation of information security has not kept pace or has only been put in place to a limited extent.
The reality behind cyber threats
Cyberattacks today are rarely spectacular. They often start small: an apparently harmless email, a reused password, or a supplier with insufficient security measures. The consequences, however, can be significant. Systems fail, service delivery comes under pressure, and executives are confronted with critical questions from supervisors, chain partners, and customers. Cybersecurity therefore is not only about technology, but above all about choices, priorities, and governance responsibility.
Guidance through models and frameworks
To bring structure to this complex landscape, many organisations make use of recognised frameworks and standards. In practice, we regularly see two models being applied.
The ISO/IEC 27001 provides a solid foundation for systematically organising information security. Not by locking everything down, but by placing risks at the centre and translating them into policies, measures, and continuous improvement. Organisations that work with ISO 27001 thereby create demonstrable control and clarity.
The NIST Cybersecurity Framework is often used as a practical growth model. Its five core functions—identify, protect, detect, respond, and recover—make cybersecurity accessible to both IT and management. The framework helps organisations gradually increase their maturity without becoming bogged down in technical details.
In practice, these frameworks complement each other well. Where ISO 27001 supports governance and assurance, NIST provides a recognisable and practical narrative for day-to-day operations..
Cybersecurity starts with awareness
What well-secured organisations have in common is not necessarily the most advanced technology, but a shared sense of awareness. Employees understand why information security matters, managers visibly take ownership, and executives ask the right questions. Cybersecurity then becomes not a barrier to innovation, but a prerequisite for safe and responsible growth.
Small steps, big impact
An effective approach does not have to be large or complex. Start by gaining insight into what is truly critical for the organisation, actively involve management, and work from a clear framework. Practise incident and crisis scenarios, evaluate (near) incidents, and continue to improve. Cyber threats are constantly evolving, and a resilient organisation evolves with them.
AI in organizations: why AI governance is no longer a luxury
Art matige intelligence (AI) is developing at lightning speed. From predictive analytics and chatbots to automated decision-making: AI is now finding its way into almost all sectors. At the same time, there is growing awareness that AI is not only a technological innovation, but also a managerial, ethical and organizational challenge. Without clear frameworks, AI can lead to risks in the areas of transparency, legality, security and trust. This is why AI governance is becoming increasingly important.
What is AI governance?
AI governance encompasses the entirety of agreements, structures, roles and processes by which organizations steer the responsible development, deployment and control of AI applications. The goal is not to inhibit innovation, but to make it verifiable, explainable and socially responsible to make.
Good AI governance connects technology with:
- strategy and policy,
- laws and regulations (such as the AI Act and AVG),
- ethical principles,
- and operational decision-making.
Why AI governance is necessary now
Many organizations are already experimenting with AI, often decentralized and pragmatic. This provides speed, but also risks:
- AI models no one knows exactly how they make decisions;
- Insufficient understanding of data quality and risk of bias;
- Unclear responsibility for errors or undesirable effects;
- Tensions between innovation, compliance and public values.
Without governance, there is a risk that AI governing organizations, rather than the other way around. It is therefore desirable that AI be approached from a policy perspective so that AI can be used effectively and the desired results can be achieved with it.
The building blocks of effective AI governance
A mature AI governance approach typically consists of the following interrelated elements:
1. Strategic framework
AI must be explicitly tied to organizational goals. What problems do we want to solve? Where do we add value - and where exactly do we not?
2. Clear roles and responsibilities.
Who owns AI applications? Who reviews for risk, ethics and legality? Consider roles such as AI owner, data owner, compliance and oversight.
3. Risk and impact analysis
Not every AI application is the same. By classifying applications (low, medium, high risk), governance can be designed proportionately.
4. Transparency and explainability
Decisions supported by or with AI must be explainable to professionals, administrators and citizens. This requires documentation, logging and understandable models.
5. Continuous monitoring and evaluation
AI is not a one-time project. Models change, data shifts and contexts evolve. Governance requires constant review and adjustment.
AI governance is also a change issue
A common misconception is that AI governance is primarily a legal or technical exercise. In practice, it is primarily about decision-making and culture. How do teams deal with uncertainty? How are assumptions made explicit? And how are different perspectives, such as legal, ethical, technical and operational integrated?
Organizations that succeed in this are not only using AI smarter, but also more carefully.
AI offers great opportunities, but only when organizations consciously manage responsible use. AI governance is the foundation for this: it ensures grip, trust and sustainable innovation.
Not by tightly regulating AI, but by combine clear frameworks with professional space.
Do you want to take steps towards mature AI governance as an organization? Continuous Connect assists in establishing governance structures, conducting AI risk assessments and effectively deploying AI within your organization.
Working data-driven: from having data to utilizing data
Many organizations today have vast amounts of data at their disposal. Yet in practice, data does not always lead to better decisions. Reports are made, dashboards are viewed, but the impact on policy, implementation and management often remains limited. Data-driven work is therefore not primarily about technology, but about How organizations use data in decision-making, policymaking or daily operations.
Why data-driven work often falters
We see similar bottlenecks recurring in many organizations:
Data is fragmented across systems and departments;
Concepts are interpreted differently or definitions differ;
Dashboards show numbers, but no action perspective;
Decisions are still made based on hierarchy or urgency.
As a result, data is primarily subsequent accountability supports, rather than advance steering.
The connection between data-driven work and governance
Effective data-driven work requires clear agreements. Without governance, there is noise, distrust and differences in interpretation. That is why data governance an indispensable prerequisite.
Key elements in this are:
1. Unambiguous definitions and terms.
Without a shared framework of concepts, departments talk past each other. A well-organized data dictionary or data catalog is the foundation here.
2. Ownership of data
Who is responsible for the quality, timeliness and meaning of data? Data-driven work requires explicit data ownership.
3. Linkage to decision-making
Data must connect to concrete decisions: in management meetings, projects and implementation. Without this connection, data degenerates into background information.
4. Transparency about assumptions and uncertainties
Data is never completely objective. Making assumptions, definitions and uncertainties explicit increases the quality of decisions.
Data-driven work is also a culture change
As with AI applications, data-driven work is not just a technical or analytical task. It requires a culture in which:
asking questions is more important than being right,
figures are cause for dialogue, not reckoning,
and professionals are supported to interpret and apply data.
Organizations that invest in this see that data-driven work leads to better prioritization, more predictability and stronger learning ability.
Data-driven work, moreover, constitutes the basis for responsible deployment of AI. Without good data, clear definitions and governance, AI is simply not reliable. In this sense, data-driven work is not an alternative to AI, but a necessary condition.
Data-driven work is not about more data, but about better decision-making. Organizations that connect data, governance and culture create calm, coherence and strategic focus. It also allows organizations to become interoperable and share data with each other across the industry.
Continuous Connect supports organizations in setting up data-driven work: from data governance, concept frameworks to decision-making structures where data actually makes a difference.
AI requires transformation
Every organization is talking about AI. Pilots, tools, experiments. But the real question is not what you can do with AI, but whether your organization is ready for it.
Because let's be honest: AI can transform organizations at lightning speed. But that transformation is rarely in the model or tool itself.
What is AI in this context?
Artificial Intelligence (AI), and in particular large language models (LLMs), enable the understanding, generation, and accessibility of language in a way that was previously unthinkable.
It feels like a revolution. And it is.
But…
The reality: AI hallucinates
AI provides answers that sound convincing but are not always correct. This is because AI does not have an understanding of truth or the real world as we know and understand it, but operates based on probability.
For organizations that rely on reliable knowledge, this is a fundamental problem.
The reflex: Deploying RAG
Many organizations are therefore focusing on RAG (Retrieval Augmented Generation): AI linked to their own data sources, so that answers are based on verified information from their own context.
That is a logical step. But it is not a quick fix.
AI transformation affects your foundation
RAG only works well if your information is in order. And that is precisely where the real transformation lies.
AI forces organizations to make their knowledge machine-readable.
That means:
- from loose documents to structured data
- from implicit knowledge to explicit definitions
- from isolated information to connected information (data + context)
In other words: from information management to organizing data.
The real transformation
AI transformation is therefore not an IT project. It is a fundamental change in how you organize, manage, and unlock knowledge, information, and data.
Organizations that understand this:
- build consistent concepts and standards
- structure their data and make it reusable
- ensure reliable sources that AI can use
Organizations that do not do this will remain dependent on generic AI with all the associated risks.
Finally
AI is advancing rapidly. But the question is not whether you use AI.
The question is:
is your information ready for AI?