---
title: Summary of McKinsey & Company’s “The State of AI:
url: https://agile-agilist.com/summary-of-mckinsey-companys-the-state-of-ai/
---

# Summary of McKinsey & Company’s “The State of AI:

How organizations are rewiring to capture value” (March 2025) — with key findings, implications for leaders (especially for your work in enterprise transformation and innovation), and coaching-led reflections.

---

### Key Findings

- **AI adoption is widespread—but value capture is still early**

  - 78 % of respondents report their organisations use AI in at least one business function (up from 72 % in early 2024).
  - Generative AI (gen AI) usage has surged: 71 % say their organisation regularly uses gen AI in at least one function, up from 65 % in early 2024.
  - Yet, fewer than 20 % of organisations say they track KPIs for their gen AI solutions; less than a third report following most of the “12 adoption-and-scaling best practices.”
  - Enterprise-wide bottom-line impact remains limited: only ~17 % of respondents say that 5 % or more of EBIT in the past 12 months is attributable to gen AI use.
- **The differentiator: structural rewiring over just tools**

  - Among 25 organisational attributes tested, **fundamental workflow redesign**  (i.e., changing how work gets done, not just automating old processes) has the biggest correlation with EBIT impact from gen AI.
  - CEO (or C-suite) oversight of AI governance is strongly associated with higher value capture; larger organisations show stronger correlation.
  - Organisations are centralising risk, compliance, and data governance elements (often via a Centre of Excellence), while tech talent and adoption tend to follow a hybrid model (part central, part distributed).
- **Workforce, talent and risk are front of mind**

  - Large organisations are more likely to have hired AI-compliance and AI-ethics specialists (13 % and 6 % of respondents respectively).
  - Reskilling of existing employees is underway: many organisations report shifting employees to new activities enabled by time saved via automation or gen AI. However, 38 % predict little or no change in workforce size over the next 3 years.
  - Risk mitigation is accelerating: 47 % of respondents say their organisations have experienced at least one negative consequence from gen AI (vs 44 % early 2024), and more organisations report actively managing risks around inaccuracy, cybersecurity, IP infringement.
- **Larger organisations (> US$500 M revenue) are moving faster**

  - These firms are more likely to follow best practices for adoption & scaling, more likely to embed gen AI across multiple functions, and more likely to report value. Smaller firms lag.

---

### Implications for Leaders & Transformation Practitioners (for your context at Agile Agilist)

- **Focus on process-and-workflow redesign** , not just model adoption. As you work with enterprises moving from Scrum to Kanban, or adjusting operating models (like your Innovation Framework or SAFe-aligned value streams), the key is *rewiring the work*  around the AI capability, not just layering AI on top of old processes.
- **Ensure executive sponsorship and governance** . In your coaching of executive programmes (SPC, ASPC), emphasise the need for top-level ownership of AI governance and strategy, not relegating AI initiatives solely to IT or digital teams.
- **Equip the workforce for transition** . As you build training for Innovation Master level or integrate digital/transformation within ASCO value streams, embed not just the tool-usage skills but the mindset shift, role redesign, and new operating rhythms that gen AI demands.
- **Embed measurement early** . One of the core gaps: tracking well-defined KPIs for AI solutions. In your consulting engagements (digital transformation, product operating model), you can help clients establish meaningful metrics (e.g., cycle-time reduction, decision-accuracy improvement, human-machine collaboration ratios) alongside adoption.
- **Manage risk consciously** . Especially for your large client base (IBM, PWC, US-AirForce etc), governance, ethics, data privacy, bias, and explainability will all be critical for scaling AI in highly regulated environments.
- **Tailor approach by size and scale** . For smaller organisations or PODs transitioning from Scrum → Kanban, the path may be more incremental; while for large enterprises, the “rewiring” must happen at scale, and they can afford to deploy more distributed models and advanced governance.

---

### Coaching Reflection

- When you coach transformation leaders, one question to ask: **“What workflow in your organisation would you *not*  worry about if gen AI solved it tomorrow?”**  If they struggle to answer, that’s a signal the organisation hasn’t yet identified the real redesign opportunity.
- Encourage leaders to shift conversation from **“Where can we plug in AI?”**  to **“How should our way of working fundamentally change because AI is now embedded?”**
- Remind them: tools matter far less than operating model. Like the electrification example we discussed, AI is not just a power-source plug but a full systems redesign.
- Use the “12 best practices” list from McKinsey (establishing a dedicated team; embedding gen AI into business processes; role-based training; tracking KPIs; fostering trust; etc.) as a coaching checklist for readiness and maturity assessment.

---

### Final Thought

The report confirms what you already emphasise in your framework: **Intentional change > technology hype.**
Adoption of AI is necessary, but **rewiring**  organisational structure, workflows, talent, governance, and metrics is what unlocks value. For your role in shaping innovation frameworks, enterprise transformations, and SAFe-aligned practices, this underscores: it’s not enough to *add AI*  — you must *embed AI thoughtfully* .

If you like, I can prepare a **slide deck summary**  of this report (3-5 slides) with clickable links and major graphs (adapted for your Innovation Master or executive-coaching toolkit). Would you like me to build that?
