The Human Algorithm

THE HUMAN ALGORITHM

AI is transforming how your company works. The question is not whether to adopt it. It is whether your culture will survive the adoption.

"The true barrier to AI adoption isn't employee readiness. It's a lack of leadership alignment on strategy, investment, and risk management."

McKinsey Global Institute, AI in the Workplace Report, 2025

Something quietly remarkable happened in most organisations over the past two years. Without a single policy mandate, without formal training programmes, and almost certainly without sign-off from the board, your employees started using AI. They drafted emails with it, summarised reports with it, wrote code with it, and started confiding in it in numbers that should give every HR leader pause. Not just for productivity tasks. Personal ones. Emotional ones. For millions of workers, an AI system has become the most psychologically safe conversation they have all day.

That is the paradox at the heart of the AI revolution in the workplace. A technology sold on efficiency is quietly reshaping something far more delicate: human connection, team cohesion, and the organisational culture that took your leadership years to build.

For CEOs, HR Directors, and Executives navigating this landscape, the challenge is no longer whether to adopt AI. That question has been answered, largely by your employees, without you. The urgent question is: how do you capture the productivity gains without losing the relational capital that actually drives performance?

Stats graphic

7.9x — More likely for employees to see AI positively impacting culture with structured adoption vs. none

81% — Of employees say AI improves their job performance

83% — Of executives agree that psychological safety measurably improves AI initiative success

1% — Of C-suite leaders describe their generative AI rollouts as "mature"

(McKinsey, 2025)

01. The Productivity Mirage

The business case for AI adoption is, on the surface, overwhelming. Industries most exposed to AI, particularly IT, financial services, and professional services, are experiencing nearly five times higher labour productivity growth than sectors slower to adopt. Nearly a third of C-suite leaders expect AI to raise revenue by more than 10% within three years.

But here is what those headlines obscure: the productivity gains are largely individual, not collective. Early AI adoption across global organisations has been concentrated almost entirely on boosting personal output, writing faster, coding faster, summarising faster. What it has not done, at least not yet, is make teams smarter, more cohesive, or more innovative together.

A landmark longitudinal study tracking a software development organisation between 2023 and 2025, one of the few long-term accounts of AI in real workplace conditions, found something counterintuitive. After two years of AI integration into daily workflows, core teamwork challenges persisted entirely unchanged. AI had reshaped individual efficiency and collaborative culture norms, but the structural problems teams faced, including miscommunication, misaligned priorities, and lack of shared understanding, remained stubbornly intact.

The implication for executives is direct: if your AI strategy is primarily a personal productivity strategy, give everyone a licence and let them figure it out, you are likely generating efficiency gains in silos while the connective tissue of your organisation quietly weakens.

"When individuals use AI in isolation, productivity gains stay isolated. When teams use AI collaboratively, sharing insights, challenging outputs, building on one another's work, the impact compounds."

Center for Creative Leadership, 2025

What the Research Actually Shows About AI & Productivity

• AI improves output on routine, well-defined tasks. The benefit decreases significantly as task complexity increases, often requiring rework that negates the gains.
• Developers with lower initial ability benefited more from AI assistance than top performers, suggesting AI can democratise workplace success, but only when people feel safe to experiment.
• AI's deeper value is the reallocation of human cognitive resources: less time on meetings and admin, more time on the creative and relational work that humans do best.

02. What AI Is Doing to Your Teams Right Now

The effects on team dynamics are not abstract. They are playing out in organisations globally, and almost certainly in yours.

Mentorship is eroding. As AI becomes the first port of call for technical questions, junior employees consult the machine rather than their senior colleagues. At Anthropic itself, arguably the organisation most aware of AI's implications, engineers noted a measurable decline in inter-colleague consultation. One engineer put it plainly: "I like working with people and it's sad that I 'need' them less now. More junior people don't come to me with questions as often." This is not just a human connection problem. It is a knowledge transfer problem. The tacit knowledge that travels between people through informal conversation, the "why" behind decisions, the institutional memory, the professional judgement honed through experience, cannot be retrieved from any AI system.

The way meetings work is shifting. With AI tools capturing and summarising meetings, teams are moving toward more asynchronous work patterns. Microsoft's 2025 Work Trend Index reports that 60% of meetings are now unscheduled, a dramatic change driven partly by AI tools that reduce the need for synchronous alignment. Self-forming, temporary project teams are replacing planned cross-functional ones. This creates agility, but it also creates fragmentation: fewer shared experiences, fewer of the spontaneous moments that build psychological safety and team identity.

Entry-level roles are disappearing. Active postings for entry-level positions are 50% lower than in 2022, a decline that correlates directly with AI implementation. Entry-level roles have traditionally served as the onboarding mechanism for organisational culture. New graduates absorb values, norms, and ways of working through proximity. Remove that pipeline and you remove a critical mechanism for cultural transmission.

The Silent Risks No One Is Talking About

When employees increasingly turn to AI for what used to be human interactions, including problem-solving, venting frustration, and seeking encouragement, three things happen gradually. Managers receive less signal about what their teams actually need. The informal relationship bonds that sustain team performance in hard times erode. And employees' capacity for interpersonal risk-taking, the foundation of psychological safety, quietly atrophies from disuse. Research from Harvard Business Review frames this starkly: if we allow AI to replace not just tasks but trust, we will see short-term gains followed by long-term erosion, rising attrition, faltering innovation, and teams that turn inward rather than toward each other.

03. Culture: The Performance Pillar That AI Can't Build

Of all the things that make organisations perform, culture is the most stubbornly human. It is not a set of values on a wall, a handbook policy, or an all-hands speech. It is the lived pattern of how we actually behave with each other when no one is watching. It is built, maintained, and transmitted through human interaction: shared experiences, honest conversations, navigated conflicts, celebrated wins, and acknowledged failures.

AI cannot do any of that. It can analyse sentiment surveys. It can flag engagement dips. It can even suggest interventions. But it cannot sit in a team debrief and model psychological safety by admitting uncertainty. It cannot read the room when someone's body language signals disengagement. It cannot carry the organisational memory of why a particular value was hard-won. And it cannot build the kind of trust that comes from being genuinely vulnerable in front of colleagues.

What AI can do, in organisations paying attention, is act as a forcing function for cultural intentionality. Organisations that were already serious about culture, that had invested in trust, psychological safety, and collaborative norms, are adopting AI faster and extracting more value from it. Those that treated culture as a secondary concern are discovering that AI adoption without cultural foundation produces compliance dressed up as progress.

A 2026 survey of over 2,800 employees found that one in three workers believes AI has negatively impacted their organisation's culture. But this effect is almost entirely driven by organisations that lacked structured adoption strategies. Where leadership takes deliberate ownership of AI integration, communicating openly, establishing governance, and training teams, the cultural effect flips dramatically positive.

"Employees in organisations with structured AI adoption are 7.9 times as likely to believe AI has positively impacted their workplace culture compared to those without formal guidance."

Perceptyx Center for Workforce Transformation, 2026

Culture Risks and Gains

Culture Risks of Unmanaged AI Adoption:

• Normalisation of individual efficiency over collective outcomes
• Erosion of mentorship and knowledge transfer pathways
• Surveillance culture when AI monitoring lacks human empathy
• Reduced psychological safety to question, experiment, or dissent
• Loss of the entry-level talent pipeline, the culture's future carriers
• Fragmentation of the shared experience that bonds team identity

Culture Gains When AI Is Integrated Intentionally:

• Real-time sentiment data enables proactive culture management
• Admin burden reduction frees managers to lead, coach, and connect
• AI democratises capability, lifting lower performers toward the mean
• Shared AI tools create new collaboration rituals and team norms
• Leadership transparency about AI builds broader organisational trust
• Cross-functional collaboration increases when AI removes process friction

04. Psychological Safety: The Non-Negotiable Foundation

If there is one concept that runs through every credible piece of research on AI and organisational performance, it is this: psychological safety determines whether AI adoption succeeds or fails.

MIT Technology Review's 2025 research found that 83% of executives believe a culture prioritising psychological safety measurably improves AI initiative success. Four in five leaders agree that organisations fostering such safety are more successful at AI adoption, and 84% have directly observed this connection in tangible outcomes.

The reason the connection is so strong is this. Rolling out AI in an organisation requires employees to do inherently risky things: admit they do not understand a tool, ask what might feel like naive questions about outputs, challenge an AI-generated recommendation, or flag a potential ethical concern about data use. In organisations where it is not safe to be uncertain, to be wrong, or to push back, employees will not do any of those things. They will nod, comply, and quietly route around the tool or use it uncritically. Neither outcome serves the organisation.

The inverse is equally powerful. In psychologically safe environments, employees experiment freely, share learnings across the team, challenge outputs rather than blindly accepting them, and collectively develop the organisational AI literacy that drives real competitive advantage. Psychological safety is not a culture nice-to-have in the AI era. It is a strategic infrastructure requirement.

The HR Imperative: What Psychological Safety Looks Like Under AI Pressure

Normalising not-knowing: Leaders who publicly share their own AI learning curves, the confusions, the failed prompts, the wrong outputs, signal that experimentation is expected, not just permitted.

• Structured space for AI critique: Team forums where AI outputs are reviewed collectively, not accepted individually, build shared critical thinking norms.
• Redesigned feedback mechanisms: Where AI has reduced informal feedback loops, formal replacements must be designed deliberately. They will not emerge on their own.
• Protected manager bandwidth: Over-spanned managers cannot build psychological safety. AI should reduce manager administrative load, not be used to justify wider spans of control.

05. The Leadership Imperative: You Are the Bottleneck

Here is the finding that should most directly concern anyone in a senior leadership role: employees are more ready for AI than their leaders. McKinsey's 2025 analysis found that the primary barrier to AI adoption is not employee resistance. It is leadership indecision, lack of vision, and failure to align on strategy. Employees are eager, often already using AI independently. Leaders are not steering fast enough.

When leadership is absent from the AI conversation, a vacuum forms. Employees fill it with their own tools, norms, and interpretations. This is why shadow AI, unsanctioned use of AI tools, is widespread in organisations without explicit strategies. It is not primarily a security risk. It is a culture risk. Uncoordinated AI use fragments teams, creates inconsistent quality standards, and prevents the collective learning that compounds over time into genuine capability.

The temptation for senior leaders is to delegate AI strategy to a technology function and monitor from a distance. The research says this reliably fails. Successful AI cultures are led from the front, by CEOs and executives who are visibly learning, experimenting, and being transparent about both the promise and the genuine uncertainty of what AI means for their organisation's future.

"AI can optimise decisions, but it can't build trust, transfer wisdom, or create connection. Those require leaders who show up for the human work."

Center for Creative Leadership, 2025

There is a concept from organisational development research that is increasingly relevant here: vertical development, the capacity to think at a higher level of systemic complexity, not just to acquire new skills. Senior leaders who approach AI only as a personal productivity tool are engaging in horizontal development: learning to use a new instrument. Vertical development asks the bigger questions. How does AI change our strategy? How does it alter who we need to be as leaders? How does it reshape the human-to-human interactions that have always been the true source of our competitive advantage?

The Five Conditions for AI-Resilient Culture

1. Strategic Clarity — Leadership communicates a clear, values-aligned AI vision. Employees with a clear leadership narrative are 1.4x more likely to trust senior management's direction.
2. Structured Adoption — Formal governance, training, and team-level AI protocols replace ad hoc individual use. The 7.9x culture benefit only occurs with structured adoption.
3. Relational Design — Intentional redesign of workflows to rebuild the human interaction that AI has optimised away, including mentorship structures, team rituals, and manager bandwidth.
4. Psychological Safety — Active investment in the conditions where teams can question AI outputs, admit uncertainty, and experiment without fear. The prerequisite for collective AI literacy.
5. Measured Culture — Track peer trust, collaboration frequency, and psychological safety with the same rigour as KPIs. What you measure signals what you value, and employees notice.

06. Where Teambuilding Becomes Strategy

There is a direct and underexamined connection between the challenges described above and the role of deliberate team development. In an era when so much workplace interaction is being mediated by AI, by asynchronous tools, and by distributed work, the experiences that bring people together in real, unstructured human interaction carry disproportionate strategic weight.

Team events and experiences, historically viewed as perks or motivational tools, are increasingly functioning as the primary mechanism through which organisations do three things AI cannot: transmit culture, build trust, and create shared identity.

The research on shared experience and team performance is long-established and consistent. Teams that have navigated challenge together, that have a shared story of collaboration, of figuring something out collectively, of having genuinely laughed and struggled and succeeded in proximity, perform differently from teams that only know each other through work outputs and screens. They communicate more honestly. They challenge each other more productively. They recover from conflict faster. They are more willing to take interpersonal risks, which is the foundation of everything that makes AI adoption work at the organisational level.

For organisations deploying AI, this reframes what team experiences are actually for. They are not rewards for performance. They are investments in the relational infrastructure that makes every other organisational initiative, including AI adoption, viable.

Specific Interventions That Build AI-Ready Culture Through Teams

• Cross-functional collaboration challenges that mirror the actual AI implementation structure, breaking the silo thinking that is the structural enemy of compounding AI gains.
• Shared challenge experiences that build the trust reserve teams draw on when AI change creates uncertainty and role ambiguity.
• Values-alignment sessions that give teams a shared language for navigating the ethical and quality questions AI tools inevitably raise.
• Leadership team offsites that address the vertical development challenge, thinking differently, not just more efficiently, about strategy in an AI-enabled organisation.
• Inclusion-focused collaboration activities that rebuild the generational and cultural connection that AI adoption can inadvertently fracture.

07. The Only Sustainable Advantage

There is a version of the AI future that companies are drifting toward without choosing it: leaner headcounts, higher individual output, frictionless processes, and organisations that feel, to the people inside them, strangely hollow. Less mentorship. Less trust. Less of the relational texture that makes coming to work feel worthwhile. Individually efficient. Collectively brittle.

There is another version, one being deliberately built by companies paying attention, where AI handles the routine, the repetitive, and the easily-summarised, freeing people to do what they have always been uniquely capable of. To connect. To challenge. To care. To build the kind of trust that survives a hard quarter, retains talent in a competitive market, and generates the creative breakthroughs that no AI system has ever produced on its own.

The organisations that will win the AI era are not those with the most advanced tools. They will be the ones with the strongest human cultures surrounding those tools, cultures built on psychological safety, relational investment, and intentional leadership. Deploy AI into a high-trust, collaborative culture and it accelerates everything good. Deploy it into a fragmented, low-trust environment and it accelerates fragmentation.

The question for every leader reading this is not whether your organisation will adopt AI. It is whether your culture is strong enough to metabolise it.

"The future of sustainable productivity isn't just AI plus human. It's AI plus human plus intentional work redesign."

Fortune, 2025

That intentional work redesign, the human side of the equation, is exactly where great teams are built, and where organisations that take culture seriously will separate from those that don't. Where good teams become great.