---
id: "lib-2026-07-09-ai-adoption-to-adaptation"
type: "library"
title: "AI adoption to adaptation"
lang: "EN"
org: "Deloitte"
source_type: "Article"
date: "2026-07-09"
week: "2026-W28"
url: "https://nosov.ai/library/lib-2026-07-09-ai-adoption-to-adaptation/"
original_url: "https://www.deloitte.com/us/en/insights/topics/talent/ai-adoption-to-ai-adaptation.html"
tier: "B"
tier_name: "Useful, limited"
tier_reason: "A Deloitte Insights practitioner article by its human capital team, built on executive interviews, company examples and Deloitte surveys and webinar polls; specific and useful, but without new controlled evidence and written by a practice that sells change programmes."
summary_basis: "full-text"
original_minutes: "14"
tags: ["enablement", "behavior"]
---

# AI adoption to adaptation

Tier B (Useful, limited): A Deloitte Insights practitioner article by its human capital team, built on executive interviews, company examples and Deloitte surveys and webinar polls; specific and useful, but without new controlled evidence and written by a practice that sells change programmes.

Summary based on the full text.

## In short

A Deloitte Insights article of 9 July 2026 by Sue Cantrell, Chloe Domergue, Allyson Dake, Jessica Murphy, Tori Sundholm and Mark Gustafson of Deloitte's human capital practice. It argues that AI adoption metrics such as active users are easy to inflate and say little about whether work has changed. The authors propose measuring "adaptation" through three behaviours: judgment about when to trust or challenge AI, experimentation, and divergent thinking. They recommend a "continuous adaptation loop" that senses behaviour from work data, intervenes with nudges, peer links or AI coaching, and designs experiences that build those behaviours over time.

## Key numbers

- 84%: organisations that have not redesigned jobs or workflows around AI (Deloitte State of AI in the Enterprise 2026)
- fewer than 60%: workers with AI access who use it in their daily workflow
- 69% / 43%: workers in heavy AI-driven change reporting higher workload / citing lack of time as the main barrier
- 78% vs 56%: employees who say they keep experimenting vs those who say they are rewarded for adapting

## What it reports

- Worker access to AI tools grew 50% in one year, yet fewer than 60% of those with access use AI daily and 84% of organisations have not redesigned jobs or workflows around it.
- Only half of executives regularly verify the quality of AI outputs when making decisions.
- In an informal Deloitte webinar poll of 1,700 respondents, judgment, divergent thinking and experimentation ranked as critical behaviours; in another of 2,422, helping workers learn to experiment ranked first among design actions.
- 78% of employees say they continually experiment with new ways of working; 56% say they are recognised or rewarded for it.
- Among workers in significant AI-driven change, 69% report higher workload and 43% cite lack of time as the biggest barrier to adapting.
- Company cases: Enel trained 1,240 managers across 18 countries over eight months on judgment and metacognition; Canva gave over 5,000 employees an AI Discovery Week; BBVA scores employee-built GPTs automatically on guardrails, clarity and context instead of reviewing each centrally; DPG Media schedules 80% of team capacity.
- A comparison of 793 sales staff at one technology company found more durable experimentation under a gamified, coached approach than under standard communications-based change.

## How they got it

A synthesis of Deloitte surveys (State of AI in the Enterprise, a February 2026 change pulse survey, Human Capital Trends), two informal webinar polls, interviews with HR and talent leaders conducted in April 2026, client work, and cited academic studies such as the Harvard and BCG "jagged frontier" field experiment.

## Limits

- Webinar polls are self-selected and informal; survey methods and samples are not detailed in the article.
- The 793-person comparison is a vendor-reported case, not an independent trial.
- Deloitte sells the change and adoption services the article recommends.

## Why it matters for companies

Many firms track AI licences and active users, and some tie adoption to performance reviews; the article shows why those numbers can rise while work stays the same. It offers concrete behavioural signals to track instead, such as whether people verify and revise outputs and how widely they use features. The workload finding is a warning for HR: time freed by AI is often absorbed or backfilled, leaving no room for the learning the tools require. The point about spans of control for people supervising multiple AI agents is a practical design question for job architecture.

Original: https://www.deloitte.com/us/en/insights/topics/talent/ai-adoption-to-ai-adaptation.html