Avinash Kaushik has a gift for shattering comfortable myths with uncomfortable data, and recently he fired a persistent theme of senior leaders pressuring their organizations to adopt AI while remaining quietly in the 1990s themselves. Notion’s data, he shared in his article, shows that most senior executives, including CEOs, are actually the most advanced AI users in the data set, operating at levels 3 and 4 at a rate six times that of individual contributors.
This inversion surprised me.
The story most practitioners tell themselves is this: AI adoption This is a top-down mandate problem where leaders demand change but are unwilling to model it. Hey, that’s what I thought too, until I read “” by Notion.Major renovation” report, a survey of more than 6,100 decision-makers and daily users of AI across 10 global markets that tells a different and more troubling story. The gap is not between the leaders pushing and the workers resisting. It is between the organizations that have moved AI from an individual tool to a system, and the overwhelming majority that have not.
This majority, by the way, is 88%. It’s bigger than a bread box, as my mother used to say.
The baseline is “early,” and this is not the exception
Notion structured its findings around a four-level maturity model. Tier 1 is AI as a thinking partner: individuals using autonomous tools to write, think, and analyze. Tier 2 as an assistant, Tier 3 as teammates, and Tier 4 is AI as a system, where autonomous agents execute complex, business-critical processes end-to-end. The distribution among 6,118 respondents: 57% at level 1, 31% at level 2, 10% at level 3 and 2% at level 4.
Twelve percent of global organizations are leveraging AI at a level where it is truly reshaping the way work gets done. Eighty-eight percent still primarily use AI in the same way you would use a better search engine.
This is especially important for Search Engine Journal readers. If you currently work in SEO or content marketing, your organization is almost certainly in this 88% group. And competitive pressure doesn’t come from organizations offering slightly better prompts. This comes from the 12% who have integrated AI into their real-world workflowsbuilt governance around this and started measuring its impact with real metrics rather than self-reported time savings.
The gap between leaders and workers is real, but management is surprising
My recent column on get AI buy-in focused on the frictions of managing change and the difficulty of moving an organization from understanding that AI research is restructuring how content is produced and measured. The Notion data adds a perspective I didn’t have then.
Decision-makers in advanced organizations describe a fundamentally different transformation than people doing their daily work. At levels 1 and 2, the arguments for AI are based almost entirely on efficiency: speed, productivity, cost reduction. At levels 3 and 4, something changes. Customer experience climbs eight percentage points as the top motivator. The activation of new abilities increases by five. Meanwhile, improving employee productivity – the dominant Tier 1-2 factor – actually loses four points among the most advanced adopters.
This is not a small reframing. It’s a completely different reason to be in the game.
For marketing teams in particular, this ties directly into what I pointed out in my column on Warning Signs Your Team Is Next to Suffer AI Removals. Teams that make the “we’re saving time” argument to justify their investment in AI are speaking Level 1-2 language. The organizations most likely to survive the next round of consolidation are those that advocate for customer experience gains and capabilities that didn’t exist before.
Why the learning curve is getting steeper, not flatter
One of the most counterintuitive findings from the Notion report is that the percentage of AI decision-makers who say investments outpace readiness is steadily increasing as organizations move forward. At level 1, 48% report this gap. At level 4, it’s 68%.
This is not what most transformation manuals predict. The hypothesis is that organizations are better at absorbing AI as they gain experience. Notion’s data conversely suggests that the more you integrate AI into real-world workflows, the harder it is for employees to keep pace with what the organization is deploying.
Singapore leads the world with 21% of organizations at level 3-4. The United States is at 11%, tied with Japan. If you work for a US company that considers itself a leader in AI, these numbers are a useful comparison with reality.
3 things that separate the 12% from the rest
The Notion data on implementation strategies is where the practitioner signal becomes clearest. Compared to Tier 1-2 organizations, advanced adopters do three things at significantly higher rates.
First of all, integration. Fifty-five percent of Tier 3-4 organizations have integrated AI into their existing systems, compared to 37% at Tiers 1-2. This 18 point gap represents the difference between AI as an add-on and AI as infrastructure. If your team continues to copy and paste results from a chat interface into your CMS or analytics platform, that’s a Tier 1 workflow, no matter how sophisticated your prompts are.
Second, governance. Forty-two percent of advanced organizations have built governance and supervisory frameworks, compared to 26% at earlier stages. This one flies in the face of how most marketing teams think about AI: governance looks like a legal department problem, not a content strategy problem. The data says otherwise. The organizations that are progressing the fastest have also progressed first in policy structures, oversight, and accountability.
Third, measurement. Thirty-seven percent of Tier 3-4 organizations measure AI impact with real-world metrics, compared to 22% at earlier stages. And these quality measures (error rates, rework) are up 19 percentage points. Workflow metrics (cycle time, throughput) increased by 15. Reported time saved – the anecdotal standard that most teams default to – is actually declining as a measurement approach among more advanced organizations.
If your organization measures AI ROI by asking people if they feel like they’re saving time, you’re measuring Tier 1 transformation with Tier 1 tools.
What Kaushik Got Right and What It Means for Your Next Team Conversation
It’s really good news that owners and managers are the most advanced AI users in the data set. Leadership behavior is one of the few reliable transmission mechanisms for organizational change. When senior executives model the advanced use of AI across a diverse set of tasks – not just writing emails, but also making decisions, executing workflows, evaluating outcomes – it creates explicit permission for the rest of the organization to take the same risks.
But there is a problem with data surfaces. This leadership intensity does not automatically translate downstream. The skills and training gap is the main challenge slowing AI adoption in Tier 3-4 organizations. The tools and role structures that incentivize senior leaders to experiment are not automatically available to individual contributors.
From what I’ve read about Notion’s data, the most dangerous position for a marketing organization right now is to be confident that they are ahead when the real benchmark is only 12% of global companies operating at the level where AI is truly reshaping manufacturing. Most teams setting aggressive AI goals are aiming for Level 2. The organizations that will matter in 18 months are those currently planning for Level 3.
Three things to bring back to your team this week for field verification. Map where your actual workflows lie in relation to Notion’s four-tier model—not where executives think they lie, but where daily work actually ends up. Identify the highest-value, recurring workflow your team performs and ask yourself if it could be automated end-to-end with human review at checkpoints rather than human execution throughout. And if you continue to measure the impact of AI by asking people if they saved time, replace that with a quality metric and a workflow metric before the next review cycle.
More resources:
Featured image: Prostock-studio/Shutterstock
