Most organizations now claim to be “using AI.” The claim is almost meaningless. McKinsey’s 2025 State of AI survey reports that 88 per cent of companies employ AI in at least one business function—yet only a third have managed to scale it across the enterprise. BCG’s parallel research is even more sobering: 60 per cent of organisations generate no material value from their AI investments. The integration model is clearly failing them.
The distinction separating the productive minority from the stalled majority is deceptively simple: are you embedding AI into how your organization thinks and operates, or bolting it onto existing processes like a chrome accessory on an aging car?
Focus On: Five Signs of Genuine Embedding
Having observed this pattern across multiple industries in previous business transformation initiatives, I’ve identified five reliable indicators that an organisation has moved beyond superficial adoption. Let me adapt them for the age of AI:
Workflows have been redesigned, not merely accelerated.
Bolt-on AI takes a broken process and makes it faster. Embedded AI prompts a more fundamental question: if we were constructing this function today, knowing what AI can do, would we design it this way at all? McKinsey’s research confirms this intuition—it identifies workflow redesign as the primary reason most companies fail to progress from pilot to scale. The organisations generating real value aren’t automating the old; they’re architecting the new.
AI budgets live in the business, not exclusively in IT.
When AI expenditure sits entirely within the technology function, it signals a bolt-on mentality. BCG’s data shows that high-performing organisations allocate AI investment across business units, with functional leaders co-owning both the spend and the outcomes. The shift is telling because AI ceases to be a technology project and becomes a business capability. In mature organisations, more than 70 per cent of staff have access to AI tools—a figure that diminishes sharply in companies where AI remains an IT initiative.
New roles exist that didn’t a year ago.
Organisations genuinely embedding AI create positions that would have been unintelligible twelve months prior: AI trainers who translate domain expertise into machine-readable formats, prompt engineers who combine linguistic precision with business acumen, human-AI facilitators who orchestrate collaboration across hybrid teams. As I examined in a previous issue on organisational design for agentic AI, these roles aren’t cosmetic additions to the org chart, rather represent a structural acknowledgment that the interface between human and artificial intelligence requires dedicated stewardship.
Governance precedes deployment, not the reverse.
Bolt-on organisations discover governance requirements after something goes wrong. Embedded organisations establish decision rights, operating boundaries, and monitoring frameworks before the first model enters production. Gartner’s warning about “agentwashing”—the tendency to rebrand simple assistants as autonomous agents—illuminates a deeper governance lacuna. Without clear taxonomies for what AI systems can decide independently, organisations accumulate technical debt dressed in transformational language.
Performance metrics have been rewritten.
Perhaps the most telling indicator. When an organisation still measures AI success through traditional KPIs—project completion rates, cost savings, processing speed—it remains in bolt-on territory. Embedded organisations track fundamentally different outcomes: how AI changes decision quality, the rate at which new capabilities emerge, whether the organisation can now answer questions it couldn’t formulate before. The shift from efficiency measurement to capability measurement is the surest sign that AI has moved from peripheral tool to operational foundation.
The Uncomfortable Implication
The data points to an uncomfortable conclusion for many executive teams. Despite headline investments—global spending exceeded $250 billion in 2024 alone—the majority of organisations have purchased proximity to intelligence without achieving integration. They own the technology but haven’t absorbed it.
The reason is rarely technical. Embedding AI demands the kind of organizational redesign that most leadership teams find genuinely disruptive: redistributing authority, retiring established metrics, creating roles that challenge existing hierarchies, and accepting that workflows perfected over decades may now represent constraints rather than assets.
Gartner projects that 40 per cent of enterprise applications will integrate task-specific AI agents by the end of 2026, up from less than 5 per cent today. That acceleration will be profoundly unforgiving to organisations still treating AI as an appendage. The gap between embedded and bolted-on won’t narrow. It will compound.
What signals have you observed that distinguish genuine AI embedding from performative adoption? And which of these indicators does your own organisation exhibit?
Follow me
That’s all for this week. To keep up with the latest in generative AI and its relevance to your digital transformation programs, follow me on LinkedIn or subscribe to this newsletter.
Disclaimer: The views and opinions expressed in Chronicles of Change and on my social media accounts are my own and do not necessarily reflect the official policy or position of S&P Global.