The Chief Data Officer who told us this was adamant about how important understanding people’s emotional responses to AI is in successfully integrating it into teams, and equally how often this is downplayed.

Leaders are facing their own AI shame, fear, or dopamine-driven responses in private, leaving their teams without role models or frameworks to help make sense of what AI means to them and how they should use it. AI use is patchy, as human-AI teams are not led or systemically enabled to create measurable business value and ROI on AI investments.

Only 37% of leaders report tangible business value from AI at scale++[CR1]++ , and only 6% report an EBIT impact of 5% or more due to AI use++[CR2]++  [1]. Efficiency, innovation, and productivity are among the promises of AI that will only be delivered when organisations create the human relationships, trust, identity, and emotional shifts that underpin high-performing Human-AI teams.

AI investment becomes measurable business value only when organisations lead AI as a human transformation, not merely a technology deployment.

The organisations pulling ahead do four things: they develop a new Human-AI leadership skill set; integrate AI as a teammate; redesign the operating system around Human-AI work; and measure leading indicators that show where value is beginning to form. Miss one factor and the others weaken. Build all four, and isolated experiments can become a repeatable source of productivity, innovation and growth.

FACTOR 1:

Two unspoken fears are eroding trust in organisations: “If I use AI, am I training my replacement?” and “If I don’t use it, will I fall behind?”[2] When leaders answer only with upbeat messages about opportunity, employees can hear toxic positivity rather than reassurance. Without acknowledging these concerns, some people freeze, some resist, and others experiment out of sight.

Leaders are facing the same fears.  Many are hesitant users who hide their knowledge gaps, particularly in the middle-management layer expected to translate strategy into everyday behaviour. Yet over 60% of leaders in organisations that are AI high performers (where AI generates +5% EBIT) role model AI usage, compared to just over 30% in other organisations. Along with workflow redesign and planning for enterprise-wide transformation, these are the top AI best practice differentiators of AI high performers [1]. People need visible permission, vision and pathways to learn, change mindsets and work differently.

Human-AI leadership is being willing to be seen as a credible learner.

A Human-AI leader speaks honestly about where AI helps, where it fails, when human judgement must overrule it and what they are still learning. That candour is what brokers trust for everyone. It turns AI from a private test of competence into a shared organisational practice.

FACTOR 2:

A global study found that 66% of workers often rely on AI without checking its accuracy, 56%have made mistakes because of unchecked recommendations [3], and 53% of C-suite respondents said they concealed their AI use to avoid judgement++++ [4]. Under those conditions, output may accelerate while judgement, quality and trust fall behind, exposing organisations to increased risk.

AI literacy is important, but generic training won’t transform a role. Teams need to understand why AI matters in their own work, which decisions it can improve and what responsible collaboration looks like. The shift is from asking, “What can the tool do for me?” to asking, “What can this Human-AI team~~~~ achieve that neither could achieve alone?”

One leader described how AI had been introduced as a productivity tool and attached to stretching targets. The team already had established processes, so the technology felt like another task bolted onto the day. A tool-led target produced resistance rather than results.

Then the leader changed their approach, instead of prescribing use, they invited the team to explore AI’s capabilities, identify valuable use cases and redesign the workflow around Human-AI collaboration. Productivity and quality improved because the work, not the tool, became the unit of change.

What teams can do now

• Discuss openly where AI is used, where it adds value and what people fear or hope it will change.
• Set Human-AI goals[CR1] , quality standards and feedback loops.
• Put quality before quantity, define where human oversight is non-negotiable.
• Onboard AI deliberately, give it context, delegate clearly, test its work and improve the collaboration over time.

FACTOR 3:

People need to feel safe enough to experiment, valued enough to share parts of their roles with AI, and ambitious enough to help shape the future. Those conditions are cultural, embedded through everyday behaviours and systems.  Culture is the human-AI infrastructure experienced through decisions, workflows, incentives, governance, and the daily signals about who is trusted to act.

Technology development is moving faster than most organisations can operationalise, which means AI strategy is happening in parallel with, or disconnected from, people, operations, product, and sales.  This looks like, ownership sitting primarily in IT and slow cross-functional decisions. Governance is built to control familiar ‘in function’ risk, rather than enable responsible enterprise-wide learning at speed.

Meanwhile, AI-ready teams pull ahead; technology leaders are creating teams of agents, while other leaders may only just be adopting AI. This is creating a widening gap in capability and opportunity, fostering competing microcultures within organisations, and weakening future talent pipelines.

The only way to create a faster organisation through AI is to create a more connected one.

A Human-AI operating model integrates technology, people, product and operational leadership. It establishes mechanisms for enterprise-wide decisions on investment and priorities. It allows intelligence to travel across boundaries, while experimentation, adaptability and collaboration become routine behaviours rather than aspirational values.  It underpins an employee experience that generates curiosity, hope and agency in creating a high-performing human-AI organisation.

What organisations can do now

• Create cultures that are curious, creative, connected and purposeful.
• Embed in organisational strategy a Human-AI operating model that’s co-led by technology, people, product and operational leaders.
• Lead the enterprise in a new era of greater interconnectedness. Reward enterprise contribution, learning and collaboration.
• Invest in data and AI infrastructure that enable intelligence to flow more freely across organisational boundaries, while also role modelling collaborative behaviours daily.
• Shift from slow governance to stewardship, so successful experiments can scale responsibly and quickly.

FACTOR 4:

What this looks like in practice

• A sales team tracks where AI supported preparation, customer engagement or deal progression, and whether that changed pipeline performance.
• A product team tracks how AI is used across routine automation, strategic thinking and ideation, then compares the quality and speed of outcomes.
• A business unit uses analytics to identify adoption hotspots, stalled workflows and opportunities to transfer successful practices.
• Leaders’ communication is reviewed for behaviours that build trust, curiosity and responsible AI use, while pulse surveys track levels of psychological safety and confidence in using AI.

These measures create short feedback loops around the behaviours and workflows that produce value. They help teams take an evidence-based approach to improving how they work with AI while it is still forming, while reinforcing the adaptive learning required in a fast-changing environment.

IN CONCLUSION:

Together, the answers to these questions prevent AI from being an isolated technology initiative. They create a platform for human-AI organisational capability that will improve productivity, accelerate innovation and create growth through human-AI thriving.

Sources:

[1] The state of AI in 2026: On the road to ROI, McKinsey Survey, August 25, 2026

[2] Psychological Safety Drives AI Adoption, Psychologytoday.com, September 22, 2025

[3] Trust, attitudes and use of artificial intelligence: A global study 2025, KPMG Report

[4] AI in the Workplace Survey, Walkme.com, August, 2025

****

Transparency note:
This article is the product of Human-AI collaboration. Microsoft Copilot was used to support research, editing and drafting. The ideas, professional judgement, interpretations and conclusions are the author's own, and responsibility for the final content rests with the author.