What Is H2H vs H2M Readiness?

In our framework, H2H concerns collaboration between people: goals, knowledge flow and psychological safety. H2M concerns AI integration into work and calibrated reliance on its output. They are complementary areas of inquiry, not a unified validated scale or proof that both must reach a particular score before AI can create value.

Organizational constraints can matter alongside model quality, data, task design and resources. This distinction helps formulate hypotheses rather than diagnose the cause of failure in advance.

What Is H2H (Human-to-Human) Readiness?

H2H readiness measures the quality of collaboration between people, before any AI is added. It rests on a core distinction: coordination is not collaboration. Coordination is simply synchronizing actions — people doing their separate tasks in sequence. Collaboration is harder: it requires goals that are genuinely compatible across people, real interdependence, and enough trust that people will disagree openly and admit mistakes.

H2H readiness is assessed across four dimensions:

• Goal alignment across levels — whether individual, team, and organizational goals are compatible rather than quietly competing.
• Boundary spanning — whether knowledge and coordination flow across functions, instead of staying trapped in silos.
• Cognitive diversity — whether different perspectives are present and actually used to reach better decisions.
• Psychological safety — whether people can surface disagreement, admit errors, and propose unconventional ideas without fear.

These dimensions can inform hypotheses alongside task, model and data evidence; no score is a universal prerequisite or guarantee of AI value.

What Is H2M (Human-to-Machine) Readiness?

H2M readiness measures how well people work with AI, not just whether they have access to it. Two organizations can use the same tools and get completely different results depending on H2M maturity.

H2M readiness is assessed across dimensions such as:

• AI integration — whether AI is built into real workflows and decision points, not bolted on as a novelty.
• Knowledge sharing between humans and machines — whether people feed AI the right context and capture what it produces back into shared knowledge.
• Calibrated trust — whether people trust AI output the right amount: neither over-relying on it where human judgment is needed, nor dismissing it where it genuinely helps.

Miscalibrated trust is a quiet value-destroyer. Over-trust leads to unchecked errors; under-trust leaves capability unused. Good H2M readiness means people know where AI helps and where their own judgment must lead.

Why Examine Both H2H and H2M?

Examine social and technical conditions together. Aligned goals and concern reporting may help, but neither guarantees value or makes model quality and data irrelevant.

An H2H-first sequence is our possible operating recommendation, not an established universal law. Task-specific experiments on AI-assisted work do not prove that every organization needs the same sequence.

Diagnosis, a pilot with a partner and measurement are independent entry points. Reuse existing evidence; the audit is not a prerequisite to a pilot.