The Human Layer of AI™
Measuring the gap between adoption and return.
Your people are already using AI. Whether that use is governed, checked and pointed at the outcomes your business is measured on is a different question, and it decides the return.
The gap is organisational
Readiness describes a stage many Gulf organisations have already passed. 93% of Middle East frontline employees use AI at least several times a week, against a 74% global average (BCG, AI at Work, 2026). Return is the open question:
Only about 6% of respondents report a significant earnings impact from AI, meaning 5% or more of EBIT (McKinsey, The State of AI: Global Survey, 2026).
84% of companies have not redesigned jobs around AI capabilities, and among workers with access to regulated tools, fewer than 60% use them in their daily workflow (Deloitte AI Institute, State of AI in the Enterprise, 2026).
56% of CEOs report no significant financial benefit from AI to date (PwC, 29th Annual Global CEO Survey, 2026).
66% of employees rely on AI output without evaluating it, and only 34% report an organisational policy or guidance on generative AI (KPMG International with the University of Melbourne, Trust, Attitudes and Use of AI, 2025; 48,340 respondents in 47 countries).
Only 28% of frontline employees see strong alignment between what their leaders say about AI and what the organisation does (BCG, AI at Work, 2026).
People
Actual use against regulated use, including unregulated tools · checking habits · capability and depth of use by role · trust and willingness to act on AI output · capacity (fatigue, change load, trust in change and AI-driven workload, measured separately).
What the Human Layer Diagnostic measures
A scored measure across three dimensions, built from a survey of leaders, a pulse of employees, leader interviews, document review and a sample of real workflows.
Strategy alignment
Sponsorship, and a written adoption and scaling plan with a named executive owner · decision rights (where AI informs, where it acts, who signs off) · AI use tied to metrics the board tracks · policy and disclosure norms known and followed · rewards and recognition that encourage AI use and redesign.
Business model
How work is organised, how teams are shaped, and what that does to cost and capacity: the extent of workflow redesign, scored from real workflows and not just from leaders' answers; team shape and roles; capacity released by AI and the share redeployed by an explicit rule.
Behaviour decides the level. A policy, strategy or training programme that exists on paper but is not reflected in what employees do does not raise a score.
The headline measure: what leaders believe against what employees report
Two gaps open every readout, reported by unit. The usage gap: leaders estimate how much AI is used, disclosed and checked, and employees report what actually happens. The say–do gap: whether leaders' messages about AI match what employees see the organisation do.
In a 2025 McKinsey study of US workplaces, 13% of employees reported using generative AI for 30% or more of their daily work; C-suite leaders estimated 4% (McKinsey, Superagency in the Workplace, 2025).
What you receive
A scored baseline across the three dimensions, by unit.
The leader–employee gap, with both components.
Three priorities, chosen by your leadership team at the readout.
A 12-month roadmap with named owners, and a briefing for each service the scores point to, with the baseline indicators as that service's success measures.
Phase | What happens | What you receive | Decision point | Duration* |
|---|---|---|---|---|
0 · Scope and sponsor | Agree on the sponsor, population, reporting units and priority workflows; data and consent terms; employee communication | Scoping note; data and consent agreement; employee communication pack | Go / no-go; sample and units confirmed | 2 weeks to 1 month |
1 · Measure | Leader survey; employee pulse; leader interviews; workflow sample; document review | Internal quality check before readout | Ready to present | 1 to 2 months |
2 · Readout and decide | Leadership readout; the team chooses what to fix first | Scored baseline report; leader–employee gap; three priorities; 12-month roadmap with named owners | What to fix first; which services to engage | up to 2 weeks |
3 · Handoff | Brief the workflow or workflows the scores point to | Handoff brief per workflow; indicator sheet | Scope and timing of follow-on work | up to 2 weeks |
4 · Re-measure | Re-run the same diagnostic on the same version | Re-score report: movement on every indicator | Continue, adjust or stop | Every 6 months |
* Duration depends on the company size and project scope
Why we re-measure
A baseline goes stale. In PwC's August 2026 snapshot of 351 CEOs, 51% had seen AI's business impact move between positive, no change and negative over eight months (PwC, CEO Survey Snapshot, 2026; a small sample, shown as a direction of travel). Gartner found that organisations that adapt their change plans based on how employees respond are four times more likely to achieve change success (Gartner, 2026; 313 senior respondents). The re-score is how you know whether the gap is closing.
What the Diagnostic does not do
It does not promise an earnings outcome. It measures the leading indicators that published research associates with AI value, and every readout says so. It does not score data quality, infrastructure or model risk. Where it finds a dependency there, it names it and says who should own it.
Who it is for
Leadership teams that have invested in AI and want to know how to maximise its value. The sponsor is typically the CEO, COO or CFO. HR is a natural partner in running it.