How to enhance your AI Maturity?
-
AI adoption has surged, making it essential for companies to assess and improve their AI maturity to stay competitive. Understanding AI maturity levels helps identify weaknesses, prioritize actions, and track progress. Tools like VTT's self-assessment or Gartner's AIMM provide frameworks to evaluate capabilities but have limitations, such as oversimplification and inconsistent benchmarks. The assessment process involves defining objectives, gathering cross-functional input, analyzing results objectively, and sharing findings, ensuring assessments drive actionable improvements. Leadership for AI maturity assessments can come from internal experts or unbiased external consultants, while strategic leaders like CTOs or CDOs typically take ownership. To improve maturity, companies should align AI goals with their business strategy, set specific actions, focus on employee engagement, and track progress. Investing in AI maturity ensures readiness for future challenges and unlocks organizational potential.
This is an improved version of my original Medium article Navigating AI Maturity from January 2023.
Artificial intelligence (AI) has been a hot topic in the business world for several years now. According to McKinsey’s The state of AI in 2022 — and a half decade in review, AI adoption has more than doubled since 2017, with top performers leveraging AI to pull ahead in their industries. And the adoption of AI will rise further across all industries according to Accenture. To remain competitive, company’s must find ways to integrate AI into their daily operations. But defining the right steps to establish AI in your company can be daunting.
To truly make the most of AI, it’s important for companies to determine their current state of AI maturity. By understanding your AI maturity level, you can better understand your areas for improvement, and make informed decisions about how to move forward with your AI initiatives. In this blog post, we will explore the key questions you need to answer to get started with assessing your organization’s AI maturity.
Why do you have to know your AI maturity?
What AI maturity models can I use?
How to assess your company’s AI maturity?
Who should lead, and who should own an AI maturity assessment?
How to improve AI maturity?
By understanding these key questions, you will be able to make informed decisions about how to best utilize AI within your organization. So, let’s dive in.
Why do you have to know your AI maturity?
If you don’t know at what maturity level you are, it’s challenging to take the right actions. For example, as someone that started with some AI experiments, building a data lake will not improve your capabilities by a lot. It only increases the risk, to burn a lot of money with no return.
One effective approach to improve your AI journey is by using maturity models. They can evaluate your current maturity level, help develop a future vision, and provide direction on how to achieve this vision. You can also use them as a benchmark against industry leaders or best practices (Saari et al., 2019). They provide guidance in prioritizing projects within your organization by identifying your weaknesses that need the biggest attention.
Once your company has a strategy in place to improve its AI capabilities, maturity assessments can help measure progress and keep you on track. Regularly assessments can help ensure that your strategy is effective, identify any obstacles that may arise, and enable you to make adjustments as needed. This way, you can stay focused on improving and making the most of your AI effort.
What AI maturity models can I use for my assessment?
What’s an AI maturity model?
Maturity models are tools that help organizations evaluate and improve their capabilities in a specific discipline. There are many maturity models out there. Some examples are: big data, cybersecurity, processes, and for this case, AI.
Maturity models measure the progress in their specific discipline through a series of levels. These reach typically from level 0 “Initial, ad-hoc” to level 5 “competent and optimizing”. Levels are defined by a set of rules. Rules typically cover the people, how the company does things, and what tools are used.
For each level, they provide guidance with action-oriented steps to reach the next level. This can be used for the company strategy and roadmap, to improve their capabilities.
Limitation of AI maturity models
Before you choose an AI maturity models (AIMM), it’s important to understand their limitations. Sadiq et al. (2021) conducted a review of the current literature on AI maturity models. Here are the main findings:
Lack of satisfaction with current approaches: Most research on AIMM focuses on developing new models. Which indicates a lack of satisfaction with current approaches by constantly developing new ones.
Lack of theoretical reflection: A solid theoretical foundation is needed to create a useful maturity model for practitioners and researchers. Which is currently not given.
Lack of validation: The majority of the studies on AIMM are empirical, using qualitative content analysis to design the model. Many AIMM’s lack validation of their structure and applicability (8 out of 15 studies), raising questions about their practical relevance and ability to accurately identify maturity levels.
Not covering all critical dimensions: As critical success factors for AIMM, the following are identified: data, analytics, technology and tools, intelligent automation, governance, people, and organization. However, none of the studies examined, evaluated all of these critical dimensions. Without including all the critical dimensions, the maturity model will not accurately reflect the current state of AI capabilities.
Additionally, there are some common criticisms of maturity models in general:
Oversimplification of Complex Progress: Many maturity models reduce their context to linear levels. This may fail to capture the unique paths and challenges each organization faces.
Inconsistent Benchmarks: There’s no universal standard, so models often define maturity based on arbitrary metrics or milestones.
Focus on Technology Over Strategy: Maturity models frequently prioritize technology capabilities over strategic alignment with business goals. True maturity involves not just deploying advanced tools but integrating technology in ways that drive core business value.
Underestimation of Organizational Culture and Skills: Success in AI requires cultural adaptation, a robust data-driven mindset, and upskilling, which many maturity models fail to address fully.
Risk of Over-Promising: AI maturity models can create unrealistic expectations among stakeholders, suggesting that simply moving to a “higher” stage will automatically yield returns. This may lead to pressure for premature AI implementation without the necessary infrastructure or expertise.
It’s important to keep in mind that AIMM are not one-size-fits-all solutions. While they can be useful in determining your company’s current state of AI maturity, take them with a grain of salt. They may not be 100% applicable to your specific needs and goals, and it’s essential to carefully consider which model is the best fit for your organization.
What AI maturity models can I find online?
To get you started quickly, I put together some examples, you can look into. It’s still essential to conduct your own research, to find the best model for your organization.
The Technical Research Centre of Finland (VTT) has created an online self-assessment tool that can provide insight into your company’s AI maturity.
The tool is free-of-charge and covers six key dimensions: Strategy and Management, Products and Services, Competence and Cooperation, Processes, Data, and Technology.
By completing the assessment, you will receive a graph that compares your company to others in your industry.
While the assessment is easy to complete and provides a good starting point, it’s important to note that it does not provide specific actions to take to improve your company’s AI maturity.
Kreutzer and Sirrenberg (2019) introduce two tools for evaluating the utilization of AI within your company. The 3-Horizons-Model is a framework for assessing the current state of AI implementation and identifying concrete steps for advancement. Additionally, the authors present the AI Maturity Map, which is a tool to assess the maturity level of your organization. It features multiple dimensions: marketing and sales, service delivery, production, AI budget, AI systems, AI staff, goals and strategy, and customer service. The book also provides guidance on how to establish an actionable AI strategy for your organization.
You can also look into the Gartner AIMM if you are searching for a more popular one. But you have to purchase the resource if you don’t already own a Gartner subscription.
How to assess your company’s AI maturity?
If you’ve decided to evaluate your company’s AI maturity, you may be wondering: How to measure your AI maturity? The sheer number of AI maturity models can be overwhelming. Stephen Gristock from ELIASSEN Group wrote a blog post over Agile Maturity Assessments, which contains valuable insights that can also be applied to AI maturity assessments. Below is a summary of the main takeaways:
It’s important to fully understand what you want to achieve. Are you looking to find out a baseline, or are you tracking progress over time? A more in-depth assessment is needed to establish a baseline, as it allows you to fully understand what you are working with. Gristock recommends asking yourself the following questions:
What are you attempting to examine?
What is your sample set?
Is it truly a cross-functional representation of the organization that we’re attempting to assess?
There are many tools or combinations of tools that can be used for an assessment, but for efficiency and accuracy, Gristock recommends using interview-based tools. To truly understand the context, choose assessments that tell the story you are looking for. Don’t conduct assessments simply for the sake of assessing; the results should lead to meaningful actions, and all participants should feel heard. The steps of an assessment are:
Planning: Define the scope of your project, take the ecosystem around into account and not only your organization.
Discovery: Within a given time frame, gather information from the right people.
Analyse: Create a report of your results and incorporate feedback until you have the final version. Rely on objectivity, expertise, and the framework.
Share your findings: Share your final report with all participants, regardless of their position.
Maintaining objectivity during an assessment is crucial for accurate, fair and representative evaluations. By defining references against which you are comparing your results, is one way to achieve it. This also allows you to have a clear standard and can help prevent bias or subjective interpretations of the data.
Here are some key points from Gristock’s article to keep in mind:
Do’s
Protect the data gathered during the assessment. Ensure that people being interviewed feel safe and can speak openly.
Exploring the data and not taking it at face value. This means looking deeper into the findings to truly understand what they mean for your organization.
Act upon the data found in the assessment. True success is achieved when the information gathered is used to make real and meaningful changes within your organization.
Don’ts
Don’t get too hung up on a single sample that may not represent the whole. It’s important to keep in mind the collected data may not be indicative of the entire organization.
Leaders should see a transformation as an ongoing effort and not a one-time thing. Improving organizational maturity is an ongoing process that requires commitment and dedication.
Who should lead, and who should own an AI maturity assessment?
Leader
Selecting a leader is a make or buy decision, and it depends on the resources and expertise that your organization already has. One option is to assign the task to an internal employee who has knowledge about AI and experience in maturity assessments. The advantage is the domain knowledge and that they are more cost-effective. On the other hand, it may be beneficial to bring in a consultant from an independent company. They are unbiased, bring valuable insights from other assessments and industry best practices, and are used to working with large data sets. In any case, it’s important to choose someone who can lead the assessment effectively and remain objective.
Owner
The ideal owners of an AI maturity assessment are those who have the ability to define the organization’s overall strategy, but also have a profound understanding of the possibilities and limitations of AI. In most cases, this responsibility falls on the shoulders of the Chief Technology Officer (CTO), Chief Information Officer (CIO), or the Chief Digital Officer (CDO). This ensures that the results are used to inform and guide decision-making at the highest level, and that the progress in improving the maturity will be aligned with the overall strategy and goals.
How to improve AI maturity?
Once you have completed your assessment, your next steps are as critical. The first step is to define your goal and align it with your overall business strategy. This will ensure that your efforts are directed towards achieving something that has an impact on your organization.
Once your goal is defined, create a roadmap that will guide you towards reaching your desired maturity level. This roadmap should contain specific steps and actions that need to be taken to achieve your goal.
Eby (2022) highlights the importance of employee satisfaction throughout the process of progressing in maturity. To truly understand and improve employee satisfaction, it is crucial to gather sentiment-driven data that paints a holistic and context-based picture of the employee experience.
This data will provide valuable insights into where improvements need to be made. Prioritizing these areas and providing training and professional-level certifications for affected employees will help to mature and improve your organization.
If your organization is looking to improve its AI maturity, one key aspect to focus on is developing a sophisticated data strategy. Luckily, my colleague Florian has put together a guide to help you get started on this topic:
Summary
Investing in your organizational maturity is investing in your success. With the many possibilities available through today’s technology, it’s worth taking the time and effort to conduct a maturity assessment and find ways to improve your business. By identifying and addressing your organization’s greatest challenges, you can keep your employees engaged, challenged, and motivated.
As we’ve learned throughout this article, it’s important to understand your organization’s current level of AI maturity to set up a strategy that will help your company improve in the right areas and track your progress. When assessing your company’s AI maturity, it’s essential to know what you want to achieve and to put a plan in place. To ensure an objective assessment, it’s crucial to stay unbiased throughout the process.
The decision of who should lead the AI maturity assessment is a “make or buy” decision. While an internal team member may have a good understanding of the company’s operations, an external consultant can bring valuable insights, an unbiased look and experience in assessing companies’ AI maturity.
It’s also crucial to choose the right AI maturity model for your assessment. There is no one-size-fits-all solution, and different models may be better suited for different industries or areas of focus. To improve your maturity after the assessment, it’s important to set up a roadmap with specific actions and track your progress over time.
In summary, investing in AI capabilities means being ready for the future in the fast-paced business environment. By understanding your AI maturity, putting a plan in place, staying objective, and tracking progress, your organization can unlock its full potential and achieve new heights of success.
Further Readings
The art of AI maturity by Accenture
Why Measure Maturity With Agile Assessments by Stephen Gristock from ELIASSEN Group
AI Business: Framework and Maturity Model by Peter Gentsch
Künstliche Intelligenz verstehen: Grundlagen — Use-Cases — unternehmenseigene KI-Journey (German) by Kreutzer & Sirrenberg
References
Eby, K. (2022, September 3). IT Maturity Models, Scorecards & Assessments | Smartsheet. Retrieved 12 January 2023, from https://www.smartsheet.com/content/it-maturity
Gristock, S. (n.d.). Why Measure Maturity With Agile Assessments? Retrieved 12 January 2023, from https://blog.eliassen.com/why-measure-maturity-with-agile-assessments
Klimko, G. (2001). Knowledge management and maturity models: Building common understanding. In Proceedings of the 2nd European conference on knowledge management (Vol. 2, pp. 269–278). Bled, Slovenia.
Kreutzer, R. T., & Sirrenberg, M. (2019). Künstliche Intelligenz verstehen: Grundlagen — Use-Cases — unternehmenseigene KI-Journey. Springer Fachmedien Wiesbaden. https://doi.org/10.1007/978-3-658-25561-9
Mettler, T. (2011). Maturity assessment models: A design science research approach. International Journal of Society Systems Science, 3(1/2), 81. https://doi.org/10.1504/IJSSS.2011.038934
Mohammad M, Mann R, Grigg N, Wagner JP. 2009. Selection of quality improvement initiatives: an initial conceptual model. Journal of Quality Measurement and Analysis 5(2):1–14.
Sadiq, R. B., Safie, N., Abd Rahman, A. H., & Goudarzi, S. (2021). Artificial intelligence maturity model: A systematic literature review. PeerJ Computer Science, 7, e661. https://doi.org/10.7717/peerj-cs.661
Saari, L., Kuusisto, O., & Pirttikangas, S. (2019). AI Maturity Web Tool Helps Organisations Proceed with AI. VTT Technical Research Centre of Finland. https://doi.org/10.32040/WhitePaper.2019.AIMaturity
Vivamus pellentesque vitae neque at vestibulum. Donec efficitur mollis dui vel pharetra.