Avinash Kaushik has a knack for crafting comfortable myths with uncomfortable data, and he recently debunked the persistent meme that senior leaders are pushing their organizations to adopt AI while sitting quietly in the 1990s themselves. The opinion data, which he shared in his post, shows that many seniors, including CEOs, are actually the most advanced AI users in the dataset, operating at Levels 3 and 4 that are six times the number of individual contributors.
That conversion surprised me.
The story that many tell themselves is that the adoption of AI is a problem of authority from the top when the leadership wants change but does not deserve it. Hey, that’s what I thought, too, until I read Notion’s “Big Fix” report, a survey of more than 6,100 AI decision makers and daily users in 10 global markets that tells a different and more inconclusive story. There is no gap between the pushing leaders and the opposing workers. It’s among organizations that have moved AI from individual tools to systems, and the vast majority that haven’t.
That majority, by the way, is 88%. That’s bigger than a breadbox, as my mother used to say.
The premise is ‘Original,’ and It’s No Exception
The theory organized its findings into a four-level growth model. Level 1 AI as a thinking partner – humans using autonomous tools to write, think, and analyze. Level 2 as an assistant, Level 3 as teammates and level 4 AI as a system, where autonomous agents make complex, business-critical processes complete. Distribution of 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 using AI to the extent that it is reshaping the way work is done. Eighty-eight percent still use AI the way you would use a better search engine.
This is very important to the readers of Search Engine Journal. If you work in SEO or content marketing right now, your organization is probably in that 88% group. And competitive pressure does not come from organizations with slightly better orders. That’s out of 12% that integrated AI into their actual workflow, built governance around it, and began measuring its impact with real metrics rather than self-reported time savings.
The Leader-Servant Gap Is Real, But The Direction Is Surprising
My recent column on getting AI buy-in focused on the conflict of change management and the difficulty of moving an organization from understanding how search AI is changing to redesigning how content is produced and measured. Notion data adds insight that I didn’t have at the time.
Decision makers in advanced organizations define change very differently than people doing everyday work. In levels 1 and 2, the AI case works almost entirely on efficiency: speed, productivity, cost reduction. In levels 3 and 4, something changes. Customer experience increases by eight percent as a top motivator. Empowering new abilities increases fivefold. Meanwhile, improving labor productivity – the dominant driver of Levels 1-2 – actually dropped four points among the most advanced students.
That’s no small redesign. It’s a completely different reason to be in the game.
For sales teams directly, this connects directly to what I highlighted in my column on the warning signs your team is following the AI cut. Teams making the “save time” argument to justify their AI investments speak Level 1-2 language. The organizations most likely to survive the next round of consolidation are those that compete with benefits and customer experiences that were not there before.
Why the Learning Curve is Steepening, Not Flattening
Another finding that contradicts the Notion report is that the percentage of AI decision makers who say investment exceeds readiness is rising slightly as organizations become more advanced. At level 1, 48% report that gap. At Level 4, it is 68%.
This is not what most transition playbooks predict. The assumption is that organizations are getting better at embracing AI as they gain experience. Notion’s data suggests the opposite in that the more deeply you embed AI into the actual workflow, the harder it is for employees to keep up with what the organization is using.
Singapore leads the world with 21% of organizations at Level 3-4. The United States sits at 11%, tied with Japan. If you work for an American company that considers itself an AI leader, those numbers are a useful reality check.
3 things that separate the 12% from the rest
Notion’s data on implementation strategies is where the signal of activity becomes more obvious. Compared to Level 1-2 organizations, advanced finders do three things at much higher rates.
First, integration. Fifty-five percent of Level 3-4 organizations have integrated AI into existing systems, compared to 37% at Level 1-2. That 18-point gap represents the difference between AI as a supplement and AI as an infrastructure. If your team is still copy-pasting output from a chat interface into your CMS or analytics platform, that’s a Level 1 workflow no matter how complex your content is.
Second, governance. Forty-two percent of advanced organizations have established governance and oversight structures, compared to 26% in earlier stages. This goes against the grain of how many marketing teams think about AI – management sounds like a legal department problem, not a content strategy problem. The data says otherwise. The fastest-moving organizations also moved first in policy, oversight, and accountability structures.
Third, moderation. Thirty-seven percent of Level 3-4 organizations measure the impact of AI with real metrics, compared to 22% in earlier measures. And those quality metrics (error rates, rework) increased by 19 percent. Workflow metrics (cycle time, throughput) increased by 15. Self-reported time saved — an anecdotal measure for many automation groups — is declining as a benchmark among more advanced organizations.
If your organization measures AI ROI by asking people if they feel they are saving time, you are measuring Level 1 conversions with Level 1 tools.
What Kaushik Gets Right And Says In The Next Group Discussion
It’s really good news that owners and managers are the most advanced AI users in the dataset. Behavioral leadership is one of the few reliable ways to convey organizational change. When seniors implement an advanced AI model using a diverse set of tasks – not just writing emails, but making decisions, implementing workflows, evaluating results – it creates a clear permission for the entire organization to take the same risk.
But there is a catch to the data areas. That leadership strength does not automatically translate into a river. The skills and training gap is the number one challenge holding back AI adoption in Level 3-4 organizations. The tools and role structures that senior leaders are most willing to explore are not automatically available to individual contributors.
My reading from the Notion data is the most dangerous situation for the marketing organization right now to be sure that you are ahead when the real benchmark is only 12% of the world’s companies that work at the level where AI is truly reformulating the result. Most of the teams setting aggressive AI goals are heading towards Level 2. The organizations that will matter in 18 months are the ones planning for Level 3 right now.
Three things to take back to your team this week to find a ground truth. Map where your actual work flow meets Notion’s four-level model, not where leadership believes it is, but where the day-to-day work actually is. Identify one high-value recurring workflow your team is running and ask if it could be automated end-to-end with human review at checkpoints instead of human execution throughout. And if you’re still measuring the impact of AI by asking people if they saved us time, get that one quality metric and one workflow metric back before the next review cycle.
Additional resources:
Featured Image: Prostock-studio/Shutterstock
