88% Of Companies Use AI As A Tool, Only 12% Built A System

New global research of 6,000+ professionals finds a stark divide in AI maturity, with leaders and workflows often further apart than teams realize.

Avinash Kaushik has a gift for puncturing comfortable myths with uncomfortable data, and recently he dismissed a persistent meme that senior leaders pressure their organizations to adopt AI while quietly staying in the 1990s themselves. Notion data, he shared in his post, shows that most senior people, including CEOs, are actually the most advanced AI users in the dataset, operating at Levels 3 and 4 at six times the rate of individual contributors.

That inversion surprised me.

The story most practitioners are telling themselves is that AI adoption is a top-down mandate problem where leadership demands change but won’t model it. Hey, that’s what I thought, too, until I read Notion’s “Great Renovation” report, a survey of more than 6,100 AI decision-makers and everyday users across 10 global markets that tells a different and more unsettling story. The gap isn’t between leaders who push and workers who resist. It’s between organizations that have moved AI from an individual tool to a system, and the overwhelming majority that have not.

That majority, by the way, is 88%. That’s bigger than a breadbox, as my mother used to say.

The Baseline Is ‘Early,’ And That’s Not The Exception

Notion structured its findings around a four-level maturity model. Level 1 is AI as a thought partner – individuals using standalone tools to draft, brainstorm, and analyze. Level 2 as an assistant, Level 3 as teammates and level 4 is AI as the system, where autonomous agents run complex, business-critical processes end-to-end. The distribution across 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 operating AI at the level where it actually reshapes how work gets done. Eighty-eight percent are still primarily using AI the way you’d use a better search engine.

This matters for Search Engine Journal readers in particular. If you’re working in SEO or content marketing right now, your organization is almost certainly in that 88% group. And the competitive pressure isn’t coming from organizations that have slightly better prompts. It’s coming from the 12% that have integrated AI into their actual workflows, built governance around it, and started measuring its impact with real metrics rather than self-reported time savings.

The Leader-Worker Gap Is Real, But The Direction Is Surprising

My recent column on getting AI buy-in focused on change management friction and the difficulty of moving an organization from understanding that AI search is changing to actually restructuring how content is produced and measured. The Notion data adds a perspective I didn’t have then.

Decision-makers at advanced organizations describe a fundamentally different transformation than the people doing the day-to-day work. At Levels 1 and 2, the case for AI runs almost entirely on efficiency: speed, productivity, cost reduction. At Levels 3 and 4, something shifts. Customer experience climbs eight percentage points as a top motivation. Enabling new capabilities climbs five. Meanwhile, improving employee productivity – the dominant Level 1-2 driver – actually drops four points among the most advanced adopters.

That’s not a small reframing. It’s a completely different reason for being in the game.

For marketing teams specifically, this connects directly to what I flagged in my column on the warning signs your team is next for AI cuts. Teams making the “we’re saving time” argument to justify their AI investment are speaking Level 1-2 language. The organizations most likely to survive the next round of consolidation are the ones arguing for customer experience gains and capabilities that didn’t exist before.

Why The Learning Curve Gets Steeper, Not Flatter

One of the more counterintuitive findings in the Notion report is the percentage of AI decision makers who say investment is outpacing readiness climbs steadily as organizations get more advanced. At Level 1, 48% report that gap. At Level 4, it’s 68%.

This is not what most transformation playbooks predict. The assumption is that organizations get better at absorbing AI as they gain experience. Notion’s data suggests the opposite that the more deeply you embed AI into actual workflows, the harder it is for employees to keep pace with what the organization is deploying.

Singapore leads globally at 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 Separating The 12% From Everyone Else

The Notion data on implementation strategies is where the practitioner signal gets clearest. Compared to Level 1-2 organizations, advanced adopters are doing three things at meaningfully higher rates.

First, integration. Fifty-five percent of Level 3-4 organizations have integrated AI with existing systems, versus 37% at Levels 1-2. That 18-point gap represents the difference between AI as an add-on and AI as infrastructure. If your team is still copy-pasting outputs from a chat interface into your CMS or analytics platform, that’s a Level 1 workflow regardless of how sophisticated your prompts are.

Second, governance. Forty-two percent of advanced organizations have built governance and oversight frameworks, compared to 26% at earlier stages. This one runs against the grain of how most marketing teams think about AI – governance sounds like a legal department problem, not a content strategy problem. The data says otherwise. The organizations moving fastest have also moved first on policy, oversight, and accountability structures.

Third, measurement. Thirty-seven percent of Level 3-4 organizations are measuring AI impact with real metrics, versus 22% at earlier stages. And those quality metrics (error rates, rework) are up 19 percentage points. Workflow metrics (cycle time, throughput) are up 15. Self-reported time saved – the anecdotal standard most teams default to – is actually declining as a measurement approach among the most advanced organizations.

If your organization is measuring AI ROI by asking people whether they feel like they’re saving time, you are measuring Level 1 transformation with Level 1 tools.

What Kaushik Got Right And What It Means For Your Next Team Conversation

88% Of Companies Use AI As A Tool

It’s genuinely good news that owners and executives are the most advanced AI users in the dataset. Leadership behavior is one of the few reliable transmission mechanisms for organizational change. When senior people model advanced AI use across a diverse set of tasks – not just drafting emails, but making decisions, running workflows, evaluating outputs – it creates explicit permission for the rest of the organization to take the same risks.

But there’s a catch the data surfaces. That leadership intensity doesn’t automatically translate downstream. The skills and training gap is the number-one challenge slowing AI adoption at Level 3-4 organizations. The tools and role structures that make senior leaders more willing to experiment are not automatically available to individual contributors.

My read on the Notion data is the most dangerous position for a marketing organization right now is being confident you’re ahead when the actual benchmark is only 12% of global companies operating at the level where AI genuinely reshapes output. Most teams setting aggressive AI goals are aiming at Level 2. The organizations that will matter in 18 months are the ones currently planning for Level 3.

Three things to take back to your team this week for ground-truthing. Map where your actual workflows sit against Notion’s four-level model, not where leadership believes they sit, but where the day-to-day work actually lands. Identify the single highest-value recurring workflow your team runs and ask whether it could be automated end-to-end with human review at checkpoints rather than human execution throughout. And if you’re still measuring AI impact by asking people whether they saved time, replace that with one quality metric and one workflow metric before the next review cycle. from 88% Of Companies Use AI As A Tool

• The Reality: Most companies use AI strictly as an individual or departmental tool for quick efficiency wins.
• The Tools: Teams rely heavily on consumer chatbots and “shadow AI” (unapproved tools employees bring to work).
• The Goal: Focuses on cost reduction and task automation rather than fundamental business model shifts. [3, 4, 5, 6, 7]

• The Reality: A small percentage of enterprises actually build AI into their core operations.
• The Difference: These “high performers” integrate AI through proprietary data pipelines, robust governance, and organization-wide talent strategies.
• The Outcome: They use AI for revenue growth and innovation, moving past pilot projects to scale benefits across the entire enterprise. [1, 2, 3, 4, 5, 6]