---
id: "framework-seven-imperatives"
type: "framework"
source_timestamps: ["§ Seven Imperatives for Creating Diverse Agentic Teams"]
tags: ["implementation-guide", "ai-strategy", "best-practices"]
related: ["concept-structural-ai-diversity", "action-diversify-tech-stack", "action-enrich-training-data", "action-fine-tune-internal-data", "action-implement-portfolio-governance", "action-cultural-red-teaming", "question-agentic-marketplaces"]
steps: ["\\\"Diversify the agentic 'tech stack': Mix foundation models from different vendors across reasoning", "evaluation", "and generation layers to prevent correlated errors.\\\"", "Enrich agentic training data: Utilize multi-dimensional psychometric datasets (like the Big Five Framework) and global cultural datasets (like the World Values Survey) to train models.", "\\\"Fine-tune through small-language models: Leverage internal enterprise data (HR systems", "employee surveys) to fine-tune agents to reflect the specific composition of the company's workforce.\\\"", "\\\"Train agents by work-shadowing humans: Allow agents to learn effective teamwork", "negotiation", "and consensus-building by analyzing email communications and meeting transcripts of diverse human teams.\\\"", "Implement a model portfolio governance policy: Establish board-level rules limiting the percentage of critical decisions that can rely on any single model vendor to mitigate concentration risk.", "\\\"Use cultural 'red-teaming': Expand traditional cybersecurity red-teaming to test LLMs for cultural sensitivity", "bias", "and societal impacts using multidisciplinary human or AI teams.\\\"", "\\\"Create agentic talent marketplaces: Utilize emerging platforms to 'recruit' AI agents with specific mixes of roles", "skills", "personality types", "and cultural backgrounds.\\\""]
sources: ["agentic"]
sourceVaultSlug: "hbr-seg-agentic"
originDay: 6
articleStem: "hbr-new-28-agent-teams-different-models"
sourceUrl: "https://hbr.org/2026/06/the-strongest-teams-of-ai-agents-will-be-built-using-different-models"
sourceTitle: "The Strongest Teams of AI Agents Will Be Built Using Different Models"
---
# Seven Imperatives for Creating Diverse Agentic Teams

A strategic framework for business and technology leaders to ensure their emerging agentic AI workforces possess true structural and cognitive diversity — avoiding the pitfalls of [[concept-correlated-ai-errors]] and [[concept-cosmetic-ai-diversity|cosmetic prompting]]. It operationalizes [[concept-structural-ai-diversity]].

**The seven imperatives:**

1. **Diversify the agentic 'tech stack'** — Mix foundation models from different vendors across reasoning, evaluation, and generation layers to prevent correlated errors. → [[action-diversify-tech-stack]]
2. **Enrich agentic training data** — Train on multi-dimensional psychometric datasets (the [[entity-big-five-framework|Big Five Framework]]) and global cultural datasets (the [[entity-world-values-survey|World Values Survey]]) to move beyond binary personas and mitigate WEIRD bias. → [[action-enrich-training-data]]
3. **Fine-tune through small-language models (SLMs)** — Use internal enterprise data (HR systems, employee surveys, psychometric evaluations) to fine-tune agents so they reflect the specific composition of the company's workforce. → [[action-fine-tune-internal-data]]
4. **Train agents by work-shadowing humans** — Let agents learn teamwork, negotiation, and consensus-building by analyzing the email communications and meeting transcripts of diverse human teams. *(No standalone action note; captured here.)*
5. **Implement a model portfolio governance policy** — Establish board-level rules limiting the percentage of critical decisions that can rely on any single vendor. → [[action-implement-portfolio-governance]] / [[concept-model-portfolio-governance]]
6. **Use cultural 'red-teaming'** — Expand traditional cybersecurity red-teaming to test LLMs for cultural sensitivity, bias, and societal impacts using multidisciplinary human or AI teams. → [[action-cultural-red-teaming]]
7. **Create agentic talent marketplaces** — Use emerging platforms to 'recruit' AI agents with specific mixes of roles, skills, personality types, and cultural backgrounds. *(Open questions in [[question-agentic-marketplaces]].)*

**Enrichment nuance:** Imperatives 1 and 6 (diversify stack; cultural red-teaming) enjoy **strong alignment** with current expert practice. Imperative 2 is conceptually sound but **not yet standard** and raises ethics/privacy questions when encoding psychometrics/values. Imperative 3 (fine-tuning on HR/employee data) is technically feasible but **high-risk** — consent, purpose-limitation (GDPR), and the danger of **codifying existing organizational bias** mean it requires strong legal/ethical review. Imperative 4's social-skills/negotiation framing is partly grounded (operational logs are used for copilots) but more speculative and surveillance-sensitive. Imperative 7 is **visionary but plausible** — a future scenario, not current reality.