Slack's 5 AI Personas vs. the 15 AI Usage Types

If you've read anything about AI adoption at work lately, you've probably met Slack's five AI personas. The Workforce Lab team at Slack sketched five characters you recognize instantly from your own office: the Maximalist, who uses AI constantly and tells everyone; the Underground, who uses it quietly and keeps it off the record; the Rebel, who wants nothing to do with the whole thing; the Superfan, who cheers enthusiastically without using it much; and the Observer, still watching from the doorway.

If you've ever tried to describe what's actually happening around you, those five names land like a relief. Finally, language for the thing everyone can see but nobody says out loud: people don't differ in how good they are at AI. They differ in how they relate to it.

We think that framing is exactly right. We also think five buckets is where the conversation starts, not where it ends. RELACTIS maps the same territory with the 15 AI usage types, scored across seven factors: Creation, Crossover, Autonomy, Advocacy, Systemization, Verification, and Engagement. This post is about what the extra resolution buys you, and roughly how the two maps line up.

What the five personas get right

Persona thinking beats the default alternative, which is a single dial labeled "AI proficiency." A dial invites ranking, ranking invites defensiveness, and defensive people tell you what you want to hear. A persona, by contrast, is a mirror. Nobody feels graded by being told they're an Observer.

The five personas also capture something a skills score structurally can't: identity and emotion. The Rebel isn't a Maximalist with lower test scores. They're having a different relationship with the technology entirely. And the Underground may be the most fluent user in the room; what sets them apart isn't ability, it's trust in the rules. Once you see adoption as a landscape of relationships instead of a leaderboard, your instincts change. You stop asking "how do we get everyone to level five?" and start asking "what does each of these people actually need?"

Where five buckets run out of road

Here's the catch: "what does this person need?" is exactly the question a five-way split struggles to answer, because the buckets are wide.

Take the Maximalist: the open, enthusiastic, heavy user. That sounds like one kind of person. Watch closely and it's at least three. One Maximalist invents uses nobody asked for and leaves a trail of half-documented experiments. Another isn't inventing much at all but has a gift for showing a colleague how AI helps their specific job. A third has gone unreasonably deep on a single use case and produces the best output in the building. Same bucket, same enthusiasm. Hand them the same task, and two of them will be miscast.

The Observer is even blurrier. One Observer is waiting to be shown, teach them once and they'll run it faithfully forever. Another has quietly decided AI isn't relevant to their work and is just being polite. A third is withholding judgment on principle until someone shows them the sources. The outward behavior, polite non-use, is identical. The needs are nearly opposite.

A rough crosswalk: five personas, fifteen types

Two caveats before the mapping. These are different instruments built by different teams, so any crosswalk is an approximation, not a Rosetta Stone. And people aren't points on a chart, plenty of us straddle buckets in both systems. With that said, here's roughly how the five personas spread across the fifteen types:

• The Maximalist roughly maps to the builder and spreader types, The Pioneer, The Remixer, The Artisan, The Systemizer, and The Evangelist, with The Self-Learner covering the quieter heavy users.
• The Underground is the cleanest match in the exercise: it roughly maps to The Shadow Adopter, the practitioner running ahead of company policy: fluent, effective, and off the books.
• The Rebel roughly maps to The Critic, with an important asterisk, because some skeptics have a genuinely sharp eye for flaws, and that eye pointed at output-checking is how a Critic grows into a Verifier.
• The Superfan (all enthusiasm, little hands-on time) roughly maps to The Sponsor and The Delegator: people whose contribution runs through others, via budget, air cover, or knowing exactly who to ask.
• The Observer fans out the widest, roughly covering The Quiet Abstainer, The Learner, and sometimes The Operator, who is simply waiting for something that just works.

Notice what the crosswalk itself tells you: three of the five personas each split into several types with different needs. That's not a knock on Slack's work, a five-persona sketch is built to start a conversation, and it does that beautifully. It's simply what compression does. The question is what gets lost in the compressed version, and whether it matters.

The resolution that changes what you do

It matters the moment you move from describing people to deciding something. Here are three calls you can only make with the finer map.

Spotting the pilot. A new tool needs a first user who runs high on Creation and Autonomy, someone cheerfully undeterred by rough edges, who will find the three uses nobody predicted. That's The Pioneer. Give the pilot to an Operator and you'll get a fair review of the onboarding flow and little else; give it to a Verifier and the pilot report becomes a risk assessment. From five buckets, all you know is who's enthusiastic. The factors tell you who's exploratory.

Spotting the quality layer. Every AI workflow eventually needs someone whose first question is "what's the source?" From the outside, that person can look like a Rebel, a brake, a no-sayer. The factor scores tell the difference: The Verifier pairs high Verification with real Engagement, while The Critic critiques from the sidelines. Mistake your Verifier for a Rebel and you sideline exactly the person the whole effort needs most.

Spotting the spreader. Being good at AI and being good at teaching AI are different traits. That's what the Advocacy factor measures. The Evangelist translates capability into a colleague's actual job: "for your work, here's how you can use it." A Pioneer's demo impresses; an Evangelist's walkthrough is still in use a week later. Pick your internal teacher by fluency alone and you'll often pick the wrong Maximalist.

Pilot, quality layer, spreader: in the five-persona map, all three are usually filed under Maximalist, or hiding among the Observers. The point of fifteen types isn't taxonomy for its own sake. It's being able to tell these people apart before the roles get handed out, not after.

Two maps, one message

The deeper agreement matters more than the differences. Both maps insist that AI adoption is not a ladder with laggards at the bottom, and that the interesting variable isn't skill but relationship. If Slack's five personas gave you words for what you were seeing at work, the fifteen types give you the next sentence: what each person needs, who they pair well with, and where their particular pitfall lies.

The fastest way to feel the difference in resolution is to try it on yourself. Reading the five personas, you probably knew your bucket within seconds. The 24-question assessment goes further: your main type, the second and third faces hiding behind it, and a seven-factor radar of how you actually relate to AI. Buckets are good for describing other people. Resolution is for understanding yourself.