Read runtime and advisory rules before applying this skill. Other Foreman layers and the catalog are in ../../content/, relative to this SKILL.md.
AI Inclusive Collaboration
Amazon warehouse workers, supervised by algorithms that tracked their every movement, eventually gathered outside their facilities with a rallying cry: "We are not robots." One employee distilled the grievance: "[Productivity] is all they care about. They don't care about their employees. They care more about the robots than they care about the employees." This was not a failure of technology. The algorithms performed precisely as designed. It was a failure of inclusion -- a leadership decision to optimise for efficiency while treating human workers as interchangeable with the machines monitoring them. The third pillar of AI-savvy leadership confronts this dynamic directly: when people are excluded from the design and deployment of AI systems, they resist, retreat, or revolt. The question every leader must answer is not whether their people can work with AI, but whether the people who will be most affected by AI were present when the system was designed.
The Framework
The Exclusion-Inclusion Spectrum
The framework identifies two fundamentally different models for integrating AI into organisations:
The Exclusion Model aims to minimise humans in the technical loop. Business operations and strategy are driven by a loop of computers where humans do not fit. Under this model, employees are valued solely for efficiency, pushed into rigid work structures supervised by AI, and stripped of autonomy. The consequences are predictable: demotivation, declining quality of human input, and counterproductive behaviour (Uber drivers, for instance, developed clever methods to manipulate algorithms that supervised them).
The Inclusion Model positions AI as a collaborator that assists in cognitively demanding tasks, leveraging AI's analytical strengths while preserving and valuing distinctly human capabilities -- moral reasoning, contextual judgement, empathy, and creative sense-making. Under this model, employees are involved in designing the work settings in which humans and machines collaborate.
The distinction between these models is the defining leadership choice of the AI era.
Three Benefits of Inclusive AI Adoption
The research identifies three specific outcomes that inclusive approaches produce:
1. Feeling of Control
When employees participate in AI design and deployment decisions, they perceive the technology as a tool that works for them rather than a control mechanism imposed upon them. A bank employee who was told by an algorithm that a well-qualified applicant did not qualify for a credit card -- and could offer no explanation -- exemplifies the opposite: a human stripped of agency, unable to exercise the judgement that defines professional competence.
2. Reduced Algorithm Aversion
Research shows that humans prefer to work with and receive advice from other humans, even when AI demonstrably outperforms human alternatives. This aversion is not rational; it is emotional. People feel uncomfortable with entities that show little similarity to themselves. Inclusive leaders address this by positioning themselves as mediators and facilitators in human-AI interactions, ensuring employees receive training to interact effectively with AI systems while maintaining access to human support when interactions go wrong.
3. Increased Trust
Nearly half of all employees globally do not trust AI at work. Trust cannot be manufactured through technical improvements alone -- better accuracy, better explainability. Trust requires that employees believe their leaders are competent, honest, and benevolent in their AI decisions. Inclusion builds this trust by giving employees visibility into how AI works, what it can and cannot do, and why it has been deployed.
Four Steps to Inclusive AI Adoption
The framework prescribes four concrete actions:
Step 1 -- Create Space for Social Connection
People who work extensively with AI experience loneliness and isolation. Digital life underwriters who assess insurance applications through AI-based systems never meet the applicants; they interact only with data. Leaders must counterbalance this by creating deliberate opportunities for human-to-human interaction -- social events, cross-team meetings, scheduled time away from screens.
Step 2 -- Make Tech and Non-Tech Teams Work Side by Side
Cross-discipline collaboration is the foundation of successful human-AI partnerships. Tech and business experts cannot retreat to separate spaces. Business experts explain organisational requirements; tech experts suggest appropriate AI approaches; HR familiarises everyone with the skills needed; operational staff integrate the human-AI workflow. The leader's role is to be an "analytics-to-business translator" -- someone who bridges jargon gaps and decomposes large problems into actionable components.
The cost of failure is illustrated by the case of a CEO at a global financial institution who interrupted his CTO's technology strategy presentation, demanding it be reduced to "three simple bullet points." The tech group pulled back from communicating with senior leadership entirely. The CEO's inability to translate between technical and business perspectives doomed the AI project within a year.
Step 3 -- Develop Leadership and Inclusion Skills
AI-savvy leaders must be perceived as someone who listens, who accounts for the uncertainties and discomfort employees feel, and who demonstrates humility about their own knowledge gaps. Research shows that employees are more willing to trust AI and engage with it when their leaders are seen as humble people willing to listen. Satya Nadella's practice of telling Microsoft employees that he, too, has knowledge gaps models the behaviour that makes inclusion possible.
Step 4 -- Reward Workers for Being Human
When AI and humans collaborate, clear guidelines must establish who -- or what -- receives credit. The 2022 Colorado State Fair art competition became a flashpoint when a human artist won using a generative AI tool. Critics dismissed his contribution as "pressing a few buttons." In reality, substantial human judgement, iterative prompting, and creative selection underpinned the work. Leaders must ensure employees share in the rewards created by human-AI collaboration, and that human contributions are recognised rather than diminished.
The Silo Problem
AI adoption deepens organisational silos in three ways:
- Technical expertise concentration -- only tech teams understand the AI systems, creating barriers for HR, operations, and marketing.
- Data ownership conflicts -- different departments maintain separate data repositories and resist sharing.
- Uneven AI impact -- some teams benefit from AI while others feel threatened, producing siloed behaviour where information is hoarded for self-protection.
When departments operate in isolation, data becomes fragmented, cross-functional problem-solving vanishes, and the organisation loses its ability to scale AI effectively.
Prompts
Prompt 1 -- Inclusion Audit:
"Conduct an inclusion audit of our AI adoption process. Our organisation [describe size, industry, AI maturity]. The AI systems we have deployed or are deploying include [describe]. For each system, assess: (1) Which stakeholders were present when the system was designed? (2) Which stakeholders were absent? (3) What are the likely consequences of their absence? (4) What specific steps can we take to bring excluded voices to the table, using the four-step inclusion framework?"
Prompt 2 -- Algorithm Aversion Mitigation:
"Our employees are resisting a new AI system [describe system and context]. Resistance manifests as [describe behaviours -- avoidance, workarounds, open complaints]. Using the framework on algorithm aversion and inclusion, help me diagnose the root cause and design an intervention that includes: (1) repositioning my role as mediator between employees and AI, (2) training that addresses emotional rather than just technical concerns, (3) mechanisms for employees to speak to a human when AI interactions go wrong, and (4) a timeline for building trust incrementally."
Prompt 3 -- Cross-Functional AI Team Design:
"Design an inclusive, cross-functional team structure for our AI initiative [describe project]. The team must include tech experts, domain experts, HR, and operational staff. For each role, specify: (1) what they contribute, (2) what questions they should be empowered to ask, (3) how to prevent tech experts from dominating discussions, and (4) how I as leader serve as the analytics-to-business translator that the framework describes."
Prompt 4 -- Human Reward Framework for AI Collaboration:
"Our organisation is deploying AI in [describe context] where humans and AI will produce outputs collaboratively. Help me design a reward and recognition framework that: (1) clearly attributes human contributions to AI-augmented work, (2) prevents the devaluation of human input, (3) establishes guidelines for crediting collaborative outputs, and (4) communicates to employees that their uniquely human capabilities -- judgement, creativity, empathy -- are valued more, not less, as AI becomes prevalent."
Use Cases
Early-Traction Customer Service Platform Automating the Wrong Tasks
A startup building an AI-powered customer service platform automates all routine queries, leaving human agents to handle only the most difficult and emotionally charged interactions -- complaints escalated because the AI failed. Agents report burnout, anxiety, and declining job satisfaction. Turnover reaches 40 percent. Applying the inclusion framework, the startup redesigns its human-AI allocation. Rather than giving humans only what AI cannot handle, the system routes a balanced mix of interactions to human agents, including straightforward ones that provide psychological relief between difficult cases. Agents are included in redesigning the routing logic, contributing domain knowledge about which interactions benefit most from human empathy. Within six months, turnover drops to 15 percent and customer satisfaction improves -- not because the AI got better, but because the humans were finally included in the design.
Growth-Stage Manufacturing Firm Where Silos Block AI Scaling
A manufacturing company successfully deploys AI in quality control at one plant. Attempts to scale the system across four additional plants fail because each plant maintains its own data repositories, uses different data formats, and resists sharing information with central IT. The pattern matches the silo diagnosis precisely. The COO establishes cross-plant AI working groups that include plant managers, quality engineers, IT staff, and frontline workers from each facility. She centralises data storage in a cloud-based platform with shared access while appointing data governance liaisons at each plant -- people who own the local relationship with both the data and the colleagues who generate it. Critically, she rewards the plants not for individual AI performance metrics but for contributions to the shared AI capability across the organisation.
Anti-Patterns
Deploying AI and announcing it after the fact. Presenting employees with a finished AI system and expecting adoption is the fastest route to resistance. Inclusion means involvement from the design stage, not notification after deployment.
Automating the easy work and leaving humans the hard parts. When AI handles routine tasks and humans receive only the difficult, ambiguous, or emotionally draining cases, the result is not efficiency. It is a recipe for burnout, resentment, and attrition. This pattern is documented in call centres where AI handles simple queries, leaving human workers to absorb the full emotional weight of dissatisfied customers.
Assuming better AI accuracy will solve the trust problem. Technical teams often respond to employee distrust by improving model performance. Research shows that trust in AI continues to decline even as accuracy improves. Trust is a leadership and inclusion problem, not a technical one.
Letting IT supervise human-AI collaboration instead of HR. When the integration of humans and AI is managed by IT departments rather than human resources, the implicit message is that employees and AI are interchangeable commodities. This structural choice is a signal of exclusion that employees read clearly.
Ignoring the loneliness tax. AI adoption increases the time employees spend interacting with screens and decreases the time they spend with other humans. Leaders who do not deliberately create space for social connection will lose engagement, creativity, and institutional knowledge to isolation.
By Stage
| Stage | Focus | Key Difference |
|---|---|---|
| Idea | Inclusive founding principles | Establish from day one that all stakeholders will participate in AI design decisions; build inclusion into the company's DNA |
| Validation | Participatory prototyping | Include target users and affected employees in AI system design before the technology is finalised |
| Early Traction | Cross-functional team formation | Build diverse teams that combine technical, domain, and operational expertise; prevent tech-only AI development |
| Growth | Silo prevention at scale | As AI spreads across departments, actively break down data silos and ensure cross-functional collaboration persists |
| Scale | Institutional inclusion architecture | Formalise inclusion practices -- reward frameworks, social connection infrastructure, human-AI collaboration guidelines -- into organisational policy |
At idea and validation stages, the cost of inclusion is low and the return is high. Including diverse perspectives in AI design at this point prevents costly redesigns later. A startup that includes frontline users in its AI design process from the beginning builds a product that works with human workflows rather than against them.
At growth and scale, inclusion becomes a governance challenge. The four-step framework must be institutionalised: social connection time built into schedules, cross-functional teams embedded in standard operating procedures, leadership humility modelled from the top, and human contribution recognised in performance and reward systems.
Output Template
Target Audience: Team / Leadership
markdown# Human-AI Collaboration Inclusion Assessment ## Date: [Date] ## AI System/Initiative: [Name] ### Stakeholder Presence Audit | Stakeholder Group | Present at Design | Present at Deployment | Impact Level | Inclusion Gap | |-------------------|------------------|-----------------------|-------------|---------------| | [Group] | [Yes/No] | [Yes/No] | [High/Med/Low] | [Describe] | ### Inclusion Model Assessment - Current model: [Exclusion / Partial Inclusion / Full Inclusion] - Evidence: [Specific indicators] - Target model: [Describe desired state] ### Four-Step Inclusion Plan | Step | Current State | Target State | Actions Required | |------|--------------|-------------|-----------------| | Social connection | [Assess] | [Goal] | [Specific actions] | | Cross-functional teams | [Assess] | [Goal] | [Specific actions] | | Leadership inclusion skills | [Assess] | [Goal] | [Specific actions] | | Human reward recognition | [Assess] | [Goal] | [Specific actions] | ### Silo Risk Assessment - Data sharing barriers: [List] - Cross-functional collaboration gaps: [List] - Recommended structural changes: [List] ### Trust Indicators to Monitor - Employee AI trust score (baseline): [Measure] - Algorithm aversion incidents: [Track] - Voluntary AI adoption rate: [Measure]
Related Skills
- AI Communication Culture -- Flat communication structures are the prerequisite for inclusive AI collaboration; without information flowing freely, inclusion remains performative.
- AI Learning for Leaders -- Leaders who understand AI can serve as the analytics-to-business translators that inclusive collaboration demands.
- AI Purpose-Driven Questions -- Inclusive collaboration ensures that purpose-driven questions reflect diverse stakeholder perspectives, not just leadership assumptions.
- Inclusive Leadership -- Covers daily DEI practices broadly; this skill focuses specifically on inclusion in human-AI collaboration design.
- Psychological Safety -- Edmondson's framework underpins the trust environment required for employees to voice concerns about AI without fear.
- Hackman's Enabling Conditions -- The structural conditions for team effectiveness apply with particular force to hybrid human-AI teams where role clarity and adequate resources determine collaborative success.
