Key Findings: The Convergent Leader: A new model for sustaining growth through disruption

Europe’s business leaders, like their counterparts elsewhere, are investing heavily in transformational technologies like artificial intelligence (AI) to drive growth. But whether this strategy unlocks AI’s growth dividend will depend, in large part, on whether the people whom leaders are counting on to do the automating and augmenting are empowered – and willing – to do so.

Focusing our inquiry on three illustrative markets, France, Germany, and the United Kingdom (UK), we set out to learn how organisations can achieve the technological transformation promised by AI. Organisations, the data reveal, are at an inflection point. Headlong investment in AI has ostensibly delivered them to the threshold of technological transformation. But for AI to deliver on its considerable growth potential, leaders must now prioritise how they engage, develop, and incentivise their people alongside the technology.

We see this as a momentous opportunity for large organisations: by empowering employees to be co-creators of their AI-enabled future, organisations can secure the trust, ingenuity, and diversity that unlock innovation and sustain growth.

Our findings also suggest that empowering the workforce in this transformation will require a new kind of leadership characterised by AI skills, inclusive behaviours, and a flexible mindset defined by relentless curiosity, grace under fire, and human-centred judgement.

Across the organisations we studied, we found leaders under intense pressure to make major artificial intelligence (AI) investments deliver growth, especially as market headwinds suppress it. In that push for transformation, however, the human work of leading people through change has been deprioritised. Employees are being asked to experiment, adapt, and carry transformation forward, often without direction, protection, or support. The result? A widening gap between what leaders expect from employees and what employees are willing and able to deliver as AI reshapes their work.

Only about one in three employees (35%) are enthusiastic adopters of AI tools.⁵ Roughly a third of employees (34%) fear they’ll be replaced by AI.⁶

Only one in three employees (33%) feel their organisation effectively communicates the impact of AI adoption.⁷

Employees recognise that leaders may not have all the answers they seek, but would nonetheless appreciate being told what leaders do know.

Nearly half the workforce (47%) is pessimistic that integrating AI into everyday work will make them more productive.¹⁰ Moreover, 37% of employees mistrust AI regarding data sharing.¹¹

Taken as a whole, our qualitative and quantitative findings suggest a profound disconnect between what leaders expect and what employees anticipate.

Leaders, we find, are looking to employees to drive AI adoption but are making tools and training available unevenly,¹⁴ conferring a powerful advantage on some while potentially disenfranchising others, most likely along existing lines of inequity that disadvantage women, people of colour, and other historically marginalised populations.¹⁵

Moreover, we see that leaders are not including employees in discussions about AI¹⁶ nor soliciting their input to inform decisions about it.¹⁷ In the absence of communication from the top, employees are thus left to imagine what AI’s implementation might mean for them. Consequently, employees are anxious, mistrustful, fearful, and cynical, sentiments that contribute to disengagement, damage psychological safety,¹⁸ and undercut AI adoption and innovation.

By keeping employees in the dark about their plans for AI, shifting responsibility for outcomes and risk mitigation to employees, and shortchanging talent development, leaders are backing away from the true demands of leadership. As a result, employees feel insecure to the point of withholding trust, commitment, and collaboration.

This dynamic suggests why investment in AI alone does not automatically translate into adoption and ultimately growth. But there’s another way forward, wherein leaders accelerate and more firmly direct the transformation they envision.

What kind of leadership regains the trust of employees and harnesses their ingenuity to unlock AI’s potential? Who do Europe’s leaders need to become to drive growth despite economic uncertainty, political realignment, tightening regulation, and a shrinking and aging workforce?

We find that effective leadership comprises capabilities across three domains, and have dubbed those who excel in these areas as Convergent Leaders.¹⁹ Roughly one in three leaders (37%) today, we find, demonstrates this convergence of capabilities.

In our full report, we outline the following key tenets for organisations, complete with tangible steps leaders and practitioners can take immediately, and long-term plans they can begin today.

To amplify growth, organisations must create the conditions that address the impediments we have surfaced in data and in conversations with business leaders in Europe.

1. Keep employees in the loop.
2. Make responsible AI part of everyday work.
3. Treat employees as partners in AI outcomes.
4. Rethink entry-level roles and early careers in an AI-enabled workplace.
5. Use data to understand AI’s impact on people and work.

Across Europe, AI is set to transform our lives and our work as we know it, in ways both novel and unknowable. But our findings ultimately reinforce decades of past research: the skills, behaviours, and mindsets that support marginalised groups are what everyone needs to thrive in an AI-enabled workplace. And whether AI investments deliver durable growth will depend less on the technology itself than on how leaders govern change and lead people through it.

Convergent Leadership offers a practical path forward. It does not require organisations to slow AI adoption or launch new initiatives. It requires leaders to activate capabilities many already possess, and apply them deliberately, in parallel with ongoing transformation. In doing so, leaders create the conditions that employees need to experiment, innovate, and take the collective risks that enterprise-wide transformation demands.

Europe’s leaders have a choice to make: they can continue to allow AI transformation to unfold unevenly and reactively. Or they can exercise the human-centred leadership that restores trust and inspires engagement. Either way, AI will continue to reshape work. But only by empowering the people who are doing that work – upholding Europe’s tradition of protecting human dignity – will leaders unleash the power of AI to augment human ingenuity, accelerate innovation, and sustain market growth.

© 2026 Coqual. All rights reserved. Unauthorized reproduction or transmission of any part of this publication in any form or by any means, mechanical or electronic, is prohibited. The findings, views, and recommendations expressed in Coqual reports are not prepared by, are not the responsibility of, and do not necessarily reflect the views of the sponsoring companies or other companies in the Coqual Task Force

1. Survey respondents were asked their agreement related to: ‘My organisation helps me build AI skills that are relevant to my current role’. They were asked to respond on a 5-point Likert scale from ‘strongly disagree’ to ‘strongly agree’. A chi-square test of independence explored whether this differed based on their country of work. The data showed a significant relationship, (χ² (8) = 88.50, p <.001). Upon exploring the standardised residuals, employees in Germany responded ‘agree’ (24%, standardised residual = -2.3) and ‘strongly agree’ (4%, standardised residual = -2.0) more than was expected, as well as UK employees responded ‘agree’ (37%, standardised residual= 3.5) and ‘strongly agree’ (9%, standardised residual = 2.9) more than was expected. French employees did not differ significantly from what was expected (standardised residual < 2.0) but 27% responded ‘agree’ and 5% ‘strongly agree’. For reporting, the data were dichotomised into ‘strongly disagree, disagree, neutral’ and ‘agree, strongly agree’.
2. Survey respondents were asked to state their agreement with the following statement: ‘Does your organisation have a responsible AI policy?’. Responses were provided via checkboxes, where agreement with the statement meant checking the box, while disagreement was indicated by leaving the box unchecked. 61% of respondents checked the box, thus agreeing with the statement.
3. Survey respondents were asked to respond to a series of statements on whether their organisations tracked any AI-related outcomes. Responses were provided via checkboxes, where agreement with the statement meant checking the boxes, while disagreement was indicated by leaving the boxes unchecked. 95% of leaders indicated that their organisations do not track anything.
4. Patrick E. Hopkins, ‘Coming soon: a lost generation of employee talent?’, Fortune, 27 January 2026, https://fortune.com/2026/01/27/lost-generation-of-employeetalent-ai-gen-z-hiring-nightmare/.
5. Survey respondents were asked ‘Which response best captures your own attitude towards AI in your organisation?’, with response options being: ‘I’m an enthusiastic adopter’, ‘I’m not convinced of its value’, ‘I’m concerned AI may replace me’, and ‘Other’. 35% of employees responded that they were enthusiastic adopters.
6. Survey respondents were asked to indicate if they were concerned about being replaced by AI or automation. Responses were provided via checkboxes, where agreement with the statement meant checking the box, while disagreement was indicated by leaving the box unchecked. A chi-square test of independence explored whether this fear differed based on their country ofwork. The data showed a mild but significant relationship (χ² (2) = 6.12, p = .05). Upon exploring the standardised residuals, no country differed significantly from what was expected (zs < 2.0), but 37% of French employees agreed they were concerned with being replaced by AI or automation (z = 1.1), along with 30% of German employees (z = -1.6), and 35% of UK employees (z = 0.4).
7. Survey respondents were asked ‘To wha extent do you feel your organisation effectively communicates the impact of AI adoption on employees across roles, levels, and functions?’. Responses were on a 5-point Likert scale from ‘not at all’ to ‘to a very large extent’. To reflect the belief that their organisation effectively communicates the impact of AI adoption on employees, data were aggregated across the ‘significantly’ and ‘to a very large extent’ categories. We found that 14% responded ‘not at all’, 16% responded ‘slightly’, 37% ‘moderately, 25% ‘significantly’, and 8% responded ‘to a very large extent’
8. Survey respondents were asked to indicate if ‘Leaders in my organisation remain confident and supportive when navigating disruption related to AI adoption’. Responses were provided via checkboxes, where agreement with the statement meant checking the box, while disagreement was indicated by leaving the box unchecked. Overall, 16% of employees agreed with the statement, while 84% did not agree. A chi-square test of independence explored whether this differed based on their country of work. The data showed a significant relationship (χ² (2) = 22.60, p < .001). The standardised residuals indicate that results from Germany did not significantly differ from what was expected: 14% of employees in Germany agreed (z = -1.1) and 86% disagreed (z = 0.5). In the UK, employees agreed (22%, z = 3.5) more than was expected, while 78% disagreed (z = -1.5). Finally, French employees agreed (12%, z = -2.3) less than was expected and 88% disagreed (z = -1.0).
9. Survey respondents were asked their agreement on the following statement: ‘To what extent do you feel leadership acts with integrity during times of business disruption?’. Responses were on a 5-point Likert scale from ‘not at all’ to ‘to a very large extent’. 9% responded ‘not at all’, 15% ‘slightly’, 43% ‘moderately’, 27% agreed ‘significantly’ with the statement and only 6% agreed ‘to a very large extent’. A chi-square test of independence explored whether responses to this statement differed based on respondents’ country of work. The data showed a significant relationship, (χ² (8) = 44.25, p < .001). In Germany, the standardised residuals indicate that 11% answered ‘slightly’ (z = -2.7), which differed significantly from what was expected. The rest of the results from Germany did not significantly differ from what was expected: 10% responded ‘not at all’ (z = 1.1), 46% ‘moderately’ (z = 1.0), 27% responded ‘significantly’ (z = 0.4) and only 6% responded ‘to a very large extent’ (standardised residual = -0.5). In the UK, 32% of employees answered ‘significantly’ (z = 2.4) more than was expected, and 8% also responded ‘to a very large extent’ (z = 2.1) more than was expected. 6% responded ‘not at all’ (z = -2.3), less than was expected, while other responses did not differ significantly from what was expected: 16% answered ‘slightly’ (z = 0.2), and 39% ‘moderately’ (z = -1.7). French employees answered ‘significantly’ (22%, z =-2.5) less than was expected, while 5% answered ‘to a very large extent’ (z = -1.4). 10% responded ‘not at all’ (z = 1.2), 19% ‘slightly’ (z = 2.3; more than was expected), and 45% ‘moderately’ (z= 0.6). For reporting purposes, the data were dichotomised into ‘not at all, slightly’ and ‘significantly, to a very large extent’.
10. Survey respondents were asked to indicate their level of agreement with the following statement: ‘I am optimistic that integrating AI in my everyday work will make me more productive’. Responses were made on a 5-point Likert scale from ‘strongly disagree’ to ‘strongly agree’; however, this outcome variable was later dichtomised, grouping ‘agree’ and ‘strongly agree’ with each other and ‘strongly disagree’, ‘disagree’, and ‘neutral’ together. 47% of respondents selected ‘strongly disagree’, ‘disagree’, or ‘neutral’, whilst 53% selected ‘agree’ or ‘strongly agree’.
11. Survey respondents were asked the following: ‘Do you feel safe having your data in AI systems?’. Response options were ‘yes’, ‘no’, or ‘unsure’. 37% of employees selected ‘no’, indicating they did not feel safe having their data in AI systems. A chi-square test of independence explored whether this sentiment differed based on their country of work. The data showed a significant relationship (χ² (4) = 41.52, p < .001). The standardised residuals show that results from Germany did not significantly differ fromwhat was expected: 39% of employees (z = -1.4) agreed , while 40% disagreed (z = 1.3). In the UK, employees agreed (51%, z = 3.1) more than was expected, and 27% disagreed (z = -4.1) less than was expected. Finally, French employees agreed (39%, z = -1.6) and 43% disagreed (z = 2.6) more than was expected.
12. Leaders’ commitment to inclusion was explored across a few different questions. For example, survey respondents were asked to indicate their agreement with the following: ‘Leaders are adopting AI solutions that minimise the negative impact on employees’. Responses were provided via checkboxes, where agreement with the statement meant checking the box, while disagreement meant leaving the box unchecked. Overall, 79% of employees disagree that leaders in their organisation adopt AI solutions that minimise negative impacts on employees. A chi-square test of independence explored whether this sentiment differed based on their country of work. The data showed a significant relationship (χ² (8) = 24.09, p < .001). Employees in Germany agreed (16%, z = -2.6) less than was expected, while 84% disagreed (z = 1.7). In the UK, employees agreed (28%, z = 3.4) more than was expected, and 72% disagreed (z = -1.8). Results from France did not significantly differ from what was expected: 20% ofFrench employees agreed (z = -0.7) and 80% disagreed (z = 0.4). Survey respondents were also asked their agreement with the following: ‘Leaders are taking measures to mitigate the disproportionate impact of AI on vulnerable segments of the workforce’. Responses were provided via checkboxes, where agreement with the statement meant checking the box, while disagreement meant leaving the box unchecked. 88% disagreed that leadership were taking measures to mitigate the impact of AI in this way. A chi-square test of independence explored whether this sentiment differed based on their country of work. The data showed a significant relationship (χ² (2) = 9.22, p = .01). The standardised residuals show that results from Germany did not significantly differ from what was expected: 11% of German employeesagreed (z = -0.6) and 89% disagreed (z = 0.2). In the UK, employees agreed (15%, z = 2.3) more than was expected, while 85% disagreed (z = -0.8). Like Germany, results from France did not significantly differ from what was expected: 10% of French employees agreed (z = -1.6) and 90% disagreed (z = 0.6). Finally, survey respondents were asked their level of agreement with the following: ‘My organisation is taking steps to ensure that AI does not amplify workforce inequities’. Responses were on a 5-point Likert scale from ‘strongly disagree’ to ‘strongly agree’. For reporting purposes, these were dichotomised into ‘strongly disagree, disagree, neutral’ and ‘agree, strongly agree’. 63% disagreed that their organisation was taking steps to ensure AI wouldn’t amplify inequities. A chi-square test of independence explored whether this sentiment differed based on their country of work. The data showed a significant relationship (χ² (8) = 59.65, p < .001). Upon exploring the standardised residuals, employees in Germany agreed (25%, z = -2.4) less than was expected, while 6% strongly agreed (standardised residual = -0.7). In theUK, employees responded ‘agree’ (38%, z = 3.5) more than was expected, and 8% responded ‘strongly agree’ (standardised residual = 1.2). French employees did not differ significantly from what was expected (zs < 2.0) but 28% agreed with the statement and 6% strongly agreed.
13. Survey respondents were asked if ‘Leaders in my organisation have made efforts to create a more inclusive workplace culture in the past 18 months’. Responses were provided via checkboxes, where agreement with the statement meant checking the box, while disagreement meant leaving the box unchecked. Overall, 78% of employees did not agree with the statement.
14. Survey respondents were asked their level of agreement with the following: ‘Our leadership is putting plans into place to reskill and upskill our people to be successful moving forward’. Responses were on a 5-point Likert scale from ‘strongly disagree’ to ‘strongly agree’; for reporting purposes, these were dichotomised into ‘strongly disagree, disagree, neutral’ and ‘agree, strongly agree’. 68% of leaders agreed or strongly agreed that they are putting effective plans into place to reskill/upskill employees. However, in contrast, employees were asked the extent to which they felt their organisation makes AI training/resources available to all employees, regardless of background or role. They were asked to respond on a 5-point Likert scale from ‘not at all’ to ‘to a very large extent’ and responses were dichotomised into ‘not at all, slightly, moderately’ and ‘significantly, to a very large extent’. Only 34% of employees agreed ‘significantly, to a very large extent’ that their organisation makes AI training/resources available to all.
15. Susan Gonzales, ‘AI literacy and the new Digital Divide – A Global Call for Action’, UNESCO, 6 August 2024 (last update 23 September 2025), https://www.unesco.org/ethics-ai/en/articles/ai-literacy-and-new-digital-divide-global-call-action.
16. Survey respondents were asked the extent to which they feel their organisation includes them in decisions around how AI is implemented in their team or role. Responses were on a 5-point Likert scale from ‘not at all’ to ‘to a very large extent’, and responses were dichotomised into ‘not at all, slightly, moderately’ and ‘significantly, to a very large extent’. Only 28% of employees agreed ‘significantly, to a very large extent’ that their organisation included them in AI-related decision making.
17. Survey respondents were asked the extent to which they feel their organisation encourages them to provide feedback on how AI is being implemented in their team or role. Responses were on a 5-point Likert scale from ‘not at all’ to ‘to a very large extent’, and responses were dichotomised into ‘not at all, slightly, moderately’ and ‘significantly, to a very large extent’. Only 30% of employees agreed ‘significantly, to a very large extent’ that their organisation seeks their input on how AI is being implemented in their team or role.
18. MIT Technology Review Insights, ‘Creating Psychological Safety in the AI Era’, MIT Technology Review, 16 December 2025, https://www.technologyreview.com/2025/12/16/1125899/creating-psychological-safety-in-the-ai-era/.
19. The Convergent Leader (n = 372, 37% of total leaders surveyed) group reflects a senior cohort where 40% (n = 148) are in second-level management roles, 48% (n = 180) are senior level managers, and 12% (n = 44) are CEOs. Gender is evenly distributed, with 48% (n = 179) women and 52% (n = 193) men. Two-thirds of this group identify as white (68%), while 32% identify with race or ethnicities other than white, including 19% Black, 4% MENA, 4.5% prefer not to say, 2% Eastern European, 2% Asian, and less than 1% indicating they are Latinx or multiracial. Educational attainment is high: 82% hold a university degree or above. This includes 4% with an associate’s degree, 28.5% with a bachelor’s degree, and 49.5% with an advanced degree. Additionally, 96.5% report not having a disability, 95% do not identify as neurodivergent, and 95% identify as heterosexual. Geographically, the majority are based in the United Kingdom (55%, n = 206), with 25% (n = 91) based in Germany and 20% (n = 75) in France.
20. Kathy Robotham, PhD, and Tara Van Bommel, PhD, ‘Developing Cognitive Flexibility,’ Adapt or Fail: How Managers Can Enable Everyone to Thrive at Work, Catalyst, 6 August 2025, https://www.catalyst.org/insights/2022/managers-adapt-future-of-work-report.
21. A one-way ANOVA was conducted to investigate the relationship between leadership clusters and whether leaders feel confident in their ability to keep up with AI developments and understand their implications for their business. The data indicated a significant effect (F(2, 991) = 254.05, p < .001), where Tukey post-hoc comparisons confirmed that Convergent Leaders scored higher than the other leadership clusters (ps < .001).
22. A one-way ANOVA was conducted to investigate the relationship between leadership clusters and whether leaders feel comfortable navigating and supporting their team through unfamiliar situations. The data indicated a significant effect (F(2, 991) = 153.40, p < .001), where Tukey post-hoc comparisons confirmed that Convergent Leaders scored higher than the other leadership clusters (ps < .001).
23. A series of chi-square tests of independence highlighted how Convergent Leaders reported thinking the following skills were extremely important for their future career success significantly more than mid-tier and lower-tier leaders: AI and technology literacy (54%), cybersecurity and privacy awareness (54%), creativity and imagination (53%), analytical thinking (51%), managing complexity (51%), digital ethics and responsible use (50%), building productive teams (48%), openness and flexibility (46%), self-leadership (46%), curiosity and learning (44%), leading across differences (44%), data fluency (43%), inspiring purpose (43%), and storytelling and communication (39%). All relationships were significant at the p < .001 level and standardised residuals were all ≥7.6. Crucially, when these skills were entered into a Principal Components Analysis, they all loaded onto a single factor, suggesting that Convergent Leaders did not discriminate between technical and non-technical skills in their pursuit of future skills.
24. A one-way ANOVA was conducted to investigate the relationship between leadership clusters and whether leaders prioritise human skills as much as technical skills across hiring, promotions and performance reviews. The data indicated a significant effect for hiring (F (2, 991) = 237.01, p < .001), promotions (F (2,991) = 247.04, p < .001), and performance reviews (F (2,991) = 246.32, p < .001), where Tukey post-hoc comparisons confirmed that high potential leaders scored higher than the other leadership clusters (p < .001).
25. Thirty-three variables including questions related to the importance of future skills, upskilling and development opportunities, inclusive leadership behaviours, and AIrelated skills were entered into a Principal Components Analysis. The analysis revealed four different factors, which explained 61% of the variance: 1) inclusive leadership behaviours, 2) skills development, 3) prioritising human skills in organisational processes, and 4) the perception of future skills. These four factors were entered into a K-means cluster analysis to explore patterns of leadership. The solution revealed three clusters: a convergent leadership group (n = 372, 37%), a mid-tier leadership group (n = 440, 44%), and a lower-tier leadership group (n =182, 18%). The convergent leadership group was then taken forward to explore their skillset further.
26. Survey respondents were asked a series of questions related to business outcomes their organisation has accomplished in the past 18 months. For example, outcomes include increased revenue or profit margin, improved customer satisfaction or loyalty, and enhanced operational efficiencies and reduced cost. All were asked with a ‘yes’ / ‘no’ response option. A series of chi-square tests of independence explored whether Convergent Leaders outperformed nonconvergent leaders on these outcomes. Analyses highlighted they did on: increased market share or entering new markets (χ² (2) = 43.64, p < .001), increased revenue or profit margin (χ² (2) = 47.26, p < .001), increased customer loyalty or satisfaction (χ² (2) = 71.55, p < .001), enhanced operational efficiencies and reduce costs (χ² (2) = 32.03, p < .001), reduced employee turnover or increased retention (χ² (2) = 7.66, p = .02), and the launch of new innovative products or services (χ² (2) = 43.30, p < .001). Survey respondents were also asked their level of agreement with the following statement: ‘In the last 18 months, my team’s productivity has increased’. Responses were on a 5-point Likert scale from ‘strongly disagree’ to ‘strongly agree’. Responses were dichotomised into ‘strongly disagree, disagree, neither agree nor disagree’, and ‘agree, strongly agree’. A chi-square test of independence explored whether Convergent Leaders outperformed non-Convergent Leaders on this outcome. The data indicated a significant effect, χ² (8) = 358.78, p < .001.
27. A chi-square test of independence explored whether empowered and unsupported employees reported different engagement levels. The data showed a significant relationship (χ² (4) = 118.05, p < .001). Empowered employees agreed (46%, z = 2.0) or strongly agreed (28%, z = 3.5) that they were engaged at work, and standardised residuals indicate that this level of agreement was more than was expected. Unsupported employees, on the other hand, agreed (37%, z = -2.3) or strongly agreed (16%, z = -4.1) that they were engaged at work, and standardised residuals suggest that this level of agreement was less than was expected.
28. A chi-square test of independence explored whether empowered and unsupported employees reported different attitudes towards AI. The data showed a significant relationship (χ² (3) = 172.62, p < .001). More empowered employees reported being enthusiastic AI adopters compared to unsupported employees (47% versus 18%). Upon exploring the standardised residuals, empowered employees indicated they were an ‘enthusiastic AI adopter’ (z = 6.8) more often than was expected, while unsupported employees responded they were an ‘enthusiastic AI adopter’ less often than was expected (z  = -8.0).
29. Eighteen variables including questions related to the organisational approach to AI adoption, inclusive leadership behaviours, and whether leadership are able to navigate this change successfully were entered into a Principal Components Analysis. The analysis revealed two different factors, which explained 60% of the variance: 1) organisational support, and 2) leadership support. These two factors were entered into a K-means cluster analysis to explore patterns of employees. The solution revealed two clusters: an empowered employee group (n = 1,096, 58%), and an unsupported employee group (n = 801, 42%).