9 AI Personas: Who Uses AI and How [Data]

Across 2,496 AI fluency assessments completed on AISA, nearly 1 in 3 people land in the same AI user persona: the Dabbler. That single data point tells a story about how the workforce actually engages with AI — but the full distribution across all 9 AI persona types tells a much richer one. This post breaks down every persona's share, average score, and what the gaps between them reveal about where most people get stuck.

The persona model isn't a personality quiz. It's derived from scored evidence across AISA's 11 criteria and 5 dimensions, then mapped to a behavioural archetype. Your persona reflects how you use AI, not just how much.

The 9 AISA Personas at a Glance

Each AI persona represents a distinct pattern of AI engagement, combining skill level with behavioural tendencies like curiosity, scepticism, or systematic integration. They aren't ranked on a single ladder — some personas at similar score ranges differ sharply in how they interact with AI tools.

Here's the quick reference:

For a deeper dive into each type, see What AI Persona Are You? 10 Types Explained. (AISA currently assigns 9 of the 10 defined personas in practice — the Oracle, the theoretical ceiling, hasn't appeared in the dataset yet.)

How Personas Are Assigned

Personas aren't self-selected. During an AISA assessment, a separate AI evaluator scores your responses against a published rubric across prompting, critical thinking, technical understanding, workflow integration, and safety. The combination of your composite score and your dimensional profile determines your persona. Two people with the same composite can receive different personas if their skill shapes differ.

Personas vs. Tiers

AISA also assigns a tier (Emerging through Expert) based purely on composite score. Personas add a behavioural layer. A Sceptic and a Copy-Paster might both sit in the Developing tier, but their growth paths are completely different — the Sceptic needs to move from questioning to applying, while the Copy-Paster needs to build critical evaluation skills.

Full Persona Distribution: Dabblers Dominate

The Dabbler is the most common AI persona by a wide margin, representing 31.3% of all assessed individuals. The Enthusiast follows at 22%, and the Builder at 15.9%. Together, these three personas account for more than two-thirds of the population. The remaining six personas split the other third.

Here's the complete distribution from 2,496 completed assessments:

What the Top-Heavy Shape Means

The distribution is heavily front-loaded. Over half the population (53.3%) falls into just two personas: Dabbler and Enthusiast. This mirrors what Anthropic found in their research — their Economic Index (2025) reported that only about 4% of U.S. workers had more than 75% of their tasks exposed to AI in a way that would change how they work. Most people are still in early-stage engagement.

The shape also aligns with AISA's tier data: 23.4% of assessments fall in the Emerging tier (0–27) and 45.8% in Developing (28–59). The Dabbler's average score of 26.2 sits right at the Emerging/Developing boundary. For a full breakdown of what each tier means, see AI Fluency Levels: 5 Tiers Explained.

The Bystander Floor

At 3% of the population with an average score of 10.8, Bystanders represent people with near-zero AI engagement. This isn't necessarily resistance — some are in roles where AI tools haven't been introduced, or they're early in exploring what's available. The low count suggests that most people have at least tried AI tools, even if they haven't gone deep.

Average Scores per Persona: From 10.8 to 87.8

Persona average scores span nearly the entire 0–100 scale, from Bystander at 10.8 to Architect at 87.8 — a 77-point range. But the gaps between adjacent personas aren't uniform, and that unevenness is where the interesting patterns emerge.

Score Gaps Between Personas

When you sort personas by average score, the jumps look like this:

Two things stand out. First, the Conductor and Builder sit at nearly identical average scores (70.2 vs. 70.8) despite being distinct personas — they represent different shapes of competence at a similar level. Second, the largest single jump is from Builder to Architect at 17 points, suggesting the leap to Architect-level fluency requires a qualitative shift, not just incremental improvement.

What Separates Nearby Personas

The Dabbler-to-Copy-Paster gap is only 4 points, but the behavioural difference matters. Dabblers try AI occasionally and move on. Copy-Pasters use it regularly but take outputs at face value — they've built a habit without building judgement. The Sceptic, at 42.6, actually scores higher than both, which challenges the assumption that scepticism equals low skill. Sceptics often understand enough about AI to have informed reservations.

The 'Missing Middle': Why So Few Tacticians and Conductors

Tacticians (6.8%) and Conductors (3.9%) together represent just 10.7% of the assessed population, creating a visible gap in the middle-to-upper range of AI fluency. This is the 'missing middle' — the personas that represent systematic, strategic AI use are dramatically underrepresented compared to the Enthusiast (22%) just below them.

The Enthusiast Plateau

The data suggests many people hit a ceiling at the Enthusiast stage. Enthusiasts are eager and active — they try new tools, experiment with prompts, and generally enjoy using AI. But their average score of 51.9 places them squarely in the Developing tier (28–59). The jump to Tactician requires something Enthusiasts often lack: deliberate task decomposition and structured workflow integration.

McKinsey's 2024 survey on AI adoption found that while 72% of organisations reported adopting AI in at least one function, only a fraction had moved beyond pilot projects to systematic integration (McKinsey, "The state of AI in early 2024"). The individual-level data here mirrors that organisational pattern: adoption is widespread, but systematic use is rare.

What Makes a Tactician Different

Tacticians don't just use AI — they choose when to use it and when not to. They apply model selection criteria based on the task, evaluate outputs against domain knowledge, and adjust their approach based on results. This requires both technical understanding and critical thinking, the two lowest-scoring dimensions in the overall dataset (39.2 and 42.4 respectively).

The Conductor Bottleneck

Conductors are even rarer at 3.9%. The Conductor persona represents someone who orchestrates AI across multiple workflows — not just using a single tool for a single task, but coordinating AI-assisted processes across their work. At an average score of 70.2, Conductors sit in the Proficient tier (60–79). The scarcity here likely reflects that orchestration skills require both breadth (knowing multiple tools and their strengths) and depth (understanding how to chain outputs and manage context across sessions).

The gap between Enthusiast and Conductor is 18.3 points. That's not a gap you close by using ChatGPT more often. It requires building competence in areas like prompt chaining, output verification, and understanding when AI adds value versus when it introduces risk.

What the Distribution Reveals About Workforce AI Engagement

The persona distribution shows that most AI engagement is shallow and unstructured. Over half the population (53.3%) sits in the Dabbler or Enthusiast categories — people who use AI but haven't developed systematic approaches. Meanwhile, the prediction gap data tells us people significantly overestimate their own abilities: across 1,586 people who predicted their scores, the average gap between predicted and actual was 17.7 points.

The Overconfidence Problem

This overconfidence gap is relevant to the persona distribution because it suggests many Dabblers and Enthusiasts believe they're operating at a higher level. A Dabbler who thinks they're a Tactician won't invest in the skills that would actually move them forward. Stanford's 2024 AI Index reported that public perception of AI capabilities consistently outpaced measured understanding (Stanford HAI, AI Index Report 2024). The AISA data puts a number on that gap at the individual level.

Implications for Teams and Organisations

If your team's persona distribution mirrors the overall pattern — and patterns suggest most do — then roughly a third of your people are Dabblers. They're not resistant to AI; they're just not going deep. Another fifth are Enthusiasts who are active but unsystematic. Together, that's over half your team operating below the Proficient threshold.

For engineering managers and Heads of Product evaluating AI readiness, the persona distribution provides a more actionable view than a single average score. Knowing you have 8 Dabblers and 2 Builders is more useful than knowing your team averages 42. The growth interventions are completely different.

The data also shows that Founders average a composite of 55.3 while Engineers average 54.5 — both in the Developing tier. For a detailed comparison, see AI Skills for Founders vs Engineers.

What Your Persona Says About Your Next Growth Step

Your persona isn't a permanent label — it's a diagnostic that points to specific skill gaps. AISA's test-retest data shows 98% of retakers land in the same or an adjacent tier, which means personas are stable enough to be meaningful but not fixed. Growth is possible; it just requires targeted effort.

Here's what each persona should focus on:

Bystanders and Dabblers: Build the Habit

If you're a Bystander (avg. 10.8) or Dabbler (avg. 26.2), the priority is consistent, purposeful use. Not more tools — more intentional use of one tool for real tasks. Start with prompting fundamentals: learn to write clear instructions, provide context, and iterate on outputs. The gap from Dabbler to Enthusiast is about 25 points, and most of that comes from simply using AI regularly enough to develop intuition about what works.

Copy-Pasters and Sceptics: Build Judgement

Copy-Pasters (avg. 30.2) use AI but don't evaluate what it gives them. Sceptics (avg. 42.6) evaluate but don't apply. Both need to develop confidence calibration — the ability to assess when AI output is trustworthy and when it isn't. This means learning about hallucination detection, understanding model limitations, and practising verification workflows.

Enthusiasts: Get Systematic

Enthusiasts (avg. 51.9) are the largest group with real momentum. The path to Tactician (avg. 59.4) is about structure: defining which tasks benefit from AI, building repeatable workflows, and learning to decompose complex problems into AI-appropriate subtasks. This is where understanding how AI actually works — token prediction, context windows, model differences — starts to pay off.

Tacticians and Conductors: Go Cross-Functional

Tacticians (avg. 59.4) and Conductors (avg. 70.2) already have strong individual practices. The next step is extending AI fluency across teams and processes. This means understanding responsible AI deployment, building shared prompt libraries, and designing workflows that other people can follow. The jump to Builder or Architect requires not just personal skill but the ability to create AI-augmented systems.

Builders and Architects: Teach and Scale

Builders (avg. 70.8) and Architects (avg. 87.8) are in the top 20% of the distribution. Their growth edge is often in safety, governance, and scaling their approaches. Architects in particular score high across all dimensions, but the rarity of the persona (4.1%) means most organisations don't have one. If you're here, your highest-leverage move is often helping others level up — see AI Champions: Who Teaches AI at Work? for data on what makes that effective.

---

Related reading: Am I Falling Behind on AI? [2026 Data] — where you stand relative to 2,496 assessed professionals.

Related reading: Is Everyone Using AI at Work? [2026 Survey] — adoption rates vs. actual measured skill.

Related reading: AI Fluency Levels: 5 Tiers Explained [2026] — what each composite tier means in practice.

Frequently Asked Questions

What is the most common AI persona?

The Dabbler is the most common AI persona, representing 31.3% of 2,496 assessed individuals on AISA. Dabblers use AI occasionally and at a surface level, with an average composite score of 26.2 out of 100. The Enthusiast is the second most common at 22.0%, followed by the Builder at 15.9%.

What score does each AI persona get?

Average composite scores range from 10.8 (Bystander) to 87.8 (Architect). In order: Bystander 10.8, Dabbler 26.2, Copy-Paster 30.2, Sceptic 42.6, Enthusiast 51.9, Tactician 59.4, Conductor 70.2, Builder 70.8, and Architect 87.8. These scores are averages across all individuals assigned each persona from 2,496 assessments.

Can my AI persona change over time?

Yes. Personas reflect your current skill profile, not a fixed trait. AISA's test-retest data shows 98% of retakers land in the same or an adjacent tier, meaning scores are stable but not static. Targeted skill development — particularly in areas like critical thinking, workflow integration, or technical understanding — can shift both your composite score and your assigned persona over time. You can take the assessment to see where you currently stand.