AI benchmark survey results
29% of organisations have not started with AI. What about you?
Dive into the results of Devoteam’s AI benchmark survey
Survey Methodology
This AI benchmark survey probes IT leaders on the current state of AI adoption in their organisations across EMEA.
It includes focus areas like strategic approaches, ROI, ethical considerations, automation trends, and the impact of AI on talent. Devoteam conducted the survey online in 2024, with more than 530 responses across 20+ countries and more than 20 industries. The respondents were diverse, including 25% C-level executives and 35% IT professionals. This allowed for a broad perspective on AI adoption and challenges across EMEA. The survey data was collected between September 10th until December 10th 2024.
530+
Survey submissions
Respondent job profiles
25%
C-level
25%
IT
20+
Countries
20+
Industries
Key takeaways
AI’s promise vs. reality
Defining Use Cases and Measuring Impact Remains a Key Challenge: While enthusiasm for AI is high, businesses are struggling to translate potential into tangible results.
(Gen)AI is rapidly becoming part of daily workflows
The productivity impact is still being evaluated. A significant percentage of employees are already using generative AI tools daily, showcasing rapid adoption. However, the survey indicates that the impact on productivity is not yet fully understood.
IT still plays a key role in AI implementation
However, Business units are increasingly taking the lead.
While IT remains a primary driver in many organisations, business units are increasingly becoming promoters of AI initiatives.
Bridging the AI skill gap
Especially in Business understanding and domain expertise, is crucial for success.
Technical skills alone are not enough; aligning AI solutions with specific business needs requires individuals who can bridge the gap between technology and business strategy.
AI readiness
While progress has been made in AI-ready infrastructure, foundational elements like data management and application modernisation still lag.
Advancements in cloud infrastructure and cybersecurity are positive, but critical foundations for AI implementation, such as data management and application modernisation, are still lacking in many organisations.
Survey Result Details
GenAI: a no-brainer for many
60% use (Gen)AI daily at work
Generative AI tools are rapidly becoming indispensable for many. The survey reveals that already 59% of respondents incorporate GenAI assistants into their daily workflows. This enthusiastic adoption underscores the transformative potential of GenAI in empowering employees with enhanced productivity and efficiency.
How much time per day do you spend using GenAI tools or GenAI assistants at work?
How has the introduction of GenAI tools affected your team’s productivity and workflow?
GenAI’s Potential
Waiting for the productivity payoff
The impact of GenAI on productivity is still unfolding. While 36% of respondents report significantly increased productivity, a notable 46% remain undecided on its effects. 8% haven’t yet experienced any changes due to AI.
The variance in GenAI’s impact on productivity could be due to several factors, including the learning curve associated with using new AI tools, the need for upskilling, and the time it takes for teams to discover the most effective use-cases for integrating AI into their workflows. This emphasises the importance of putting people first in the AI adoption.
Our survey reveals a critical gap: while 60% of employees use GenAI daily at work, businesses struggle to define use cases and measure impact. We help translate this enthusiasm into strategic value.
Olivier Mallet
AI Agency Director at Devoteam
AI’s promise vs. reality
Businesses struggle to define use cases and measure impact
There is a disconnect between AI’s potential and its practical implementation. While businesses recognise AI’s power, they face challenges. Which ones? Identifying concrete use cases, measuring impact, and managing the associated costs and technological changes. Defining an AI strategy with a measured approach is needed to bridge this gap. Keep in mind: build a strategy over shiny tools.
What is the most significant barrier to achieving a benefit for AI in your organisation?
How would you characterise your level of trust in the AI tools your organisation currently utilises?
Crucial for AI adoption
Trust in AI tools is rather high
The level of trust in AI tools is encouraging, with 62.63% of respondents giving a score of 4 or 5. This suggests a positive trend towards the acceptance of AI in the workplace.
However, a significant 28% of respondents remain neutral (with a score of 3), indicating a need to address concerns and build greater confidence in AI systems.
Who is the promoter of AI?
IT still leads, but Business takes the wheel
While IT departments remain the primary drivers of AI implementation in 55% of organisations, a shift is underway. 29% of surveyed companies identify business departments as the primary promoters of AI.
This suggests a growing understanding of AI’s strategic value beyond technical applications. However, other departments like Marketing, Finance, HR, and Legal remain largely on the sidelines, indicating a need to broaden awareness and explore AI’s potential across all business functions.
How would you characterise your level of trust in the AI tools your organisation currently utilises?
It is great to see that a lot of the AI drive comes from the business. It is up to the IT team to make sure the foundations are on such a level that the ROI can be accomplished from the business use cases.
Gert Jan van Halem
CTO at Devoteam
Which of the following best describes your organisation’s current state of generative AI adoption?
(Gen)AI adoption
Hesitation and early stages dominate
This data reveals a generative AI landscape characterised by both cautious exploration and nascent implementation. A significant portion of organisations (32%) have yet to initiate any generative AI efforts. While a combined group of 15% are either piloting applications (12%) or have limited production deployments (3%), the majority are still in the early stages of strategising, identifying use cases or not ready to implement use cases.
Interestingly, a notable 18% have established a formal generative AI strategy, suggesting a proactive approach. Yet only 10% have successfully implemented their first use case. This data indicates a landscape where many are laying the groundwork, but widespread adoption and integration of generative AI are still on the horizon.
Bridging the skill gap
AI skill gaps: A critical bottleneck
The data reveals a diverse range of AI skill gaps faced by organisations. ‘Business understanding and domain expertise related to AI applications’ emerges as the most critical gap (32%), highlighting the challenge of aligning AI solutions with specific business needs. This is followed by ‘AI engineering and development skills’ (17%) and ‘data science and machine learning expertise’ (15%), indicating a need for technical talent to build and deploy AI solutions.
Interestingly, ‘AI prompting’ (7%) is among the least cited skill gaps, possibly suggesting that it’s perceived as a more advanced skill or that its importance is not yet fully recognised.
The prevalence of these skill gaps underscores the need for strategic upskilling and talent acquisition initiatives to bridge the divide and enable successful AI adoption.
What are the most critical AI skill gaps you currently face in your organisation?
73% report moderate to low AI integration. We bridge this gap with robust AI service delivery, ensuring seamless implementation and measurable impact within existing IT ecosystems.
Patricia Milheiro
AI Service Director at Devoteam
Please rate your organisation’s readiness in the following areas: (Scale of 1-5, 1 being low maturity)
AI readiness
Foundations still under construction
While some areas demonstrate strong readiness, gaps persist in crucial foundational elements for AI implementation.
For the overall readiness, we see that over 20% of organisations have yet to build core foundations in application modernisation, cybersecurity, data management, and cloud infrastructure. On the other hand, we see that organizations show significant progress in AI-ready cloud infrastructure and cybersecurity.
Where can they improve? Data management and application modernisation lag behind. Less than half of the organisations can effectively leverage data for AI purposes, potentially hindering AI effectiveness.
AI integration
Rather not integrated
The integration of AI tools and platforms with existing IT infrastructure and operational systems presents a varied picture.
The integration levels range from seamless integration (8% “Very high”) to significant challenges (18% “Very low”).
A considerable portion (73%) report moderate (“Medium”) to poor (“Low” or “Very low”) integration. This indicates opportunities to improve compatibility and streamline AI tool and platforms deployment within existing technology ecosystems.
How well integrated are your AI tools and platforms with your existing IT infrastructure and operational systems?
AI monitoring capabilities
Room for improvement
A significant number of organisations have limited or basic AI monitoring capabilities. 46% report “basic monitoring with manual intervention,” while 31% have “limited or no monitoring capabilities.” This lack of mature monitoring presents potential risks, including undetected performance drift, bias, and fairness issues.
While regulations around AI monitoring are still evolving, organisations should strive for more comprehensive monitoring practices. This includes real-time insights, automated alerts, and remediation capabilities to ensure responsible and reliable AI deployments.
How mature is your organisation’s capability to monitor and observe AI models in production for performance, drift, and potential biases?
AI compliance processes
A growing awareness for AI compliance
A significant portion of organisations (32%) have yet to establish formal processes. On the other end of the spectrum, 7% have implemented rigorous processes with continuous monitoring and auditing.
The majority (61%) fall in between these extremes, with some processes in place but not fully integrated (41%) or well-defined processes integrated into AI workflows (19%). This indicates a growing awareness of the need for AI compliance but also highlights the challenge of fully embedding these processes across organisations.
Has your organisation established processes to ensure AI systems and applications comply with relevant regulations and ethical guidelines?
The right time to talk about your AI journey is right now.