Home / Resources / Hiring AI talent / AI fluency
Hiring guide · Non-engineering roles
Twelve open questions for marketing, sales, operations and support roles, a 1 to 4 rubric, and a 30-minute work sample you can run next week.
AI fluency is the ability to use AI tools well: knowing when to use them, how to brief them, how to check the output and who owns the result. To test it in non-engineering hires, ask open questions about real work, watch a short live task and score both on one rubric.
The clearest definition I have found is in the AI Fluency Framework, written by Rick Dakan and Joseph Feller with Anthropic. It defines AI fluency as interacting with AI systems in ways that are "effective, efficient, ethical and safe." It then splits that into four competencies. None of them require writing code.
| Competency | What the framework says | What it looks like in a non-engineering job |
|---|---|---|
| Delegation | Setting goals and deciding whether, when and how to use AI | Knows which parts of the job to hand to a tool and which to keep |
| Description | Describing goals well enough to get useful output | Gives the tool context, examples and a clear format, then revises |
| Discernment | Judging how useful the output is | Checks facts, numbers and tone before anything goes to a customer |
| Diligence | Taking responsibility for what you do with AI and how | Keeps private data out of public tools and owns the final result |
The middle column paraphrases the framework. The right column is my own reading of what each one looks like at work.
Most people who use AI at work are not engineers. In Microsoft and LinkedIn's 2024 Work Trend Index, a survey of 31,000 knowledge workers in 31 markets, 75% said they use AI at work. In the same report, 66% of leaders said they would not hire someone without AI skills. And 71% said they would rather hire a less experienced candidate with AI skills than a more experienced one without them.
Some companies now make it a rule. Zapier announced in May 2025 that all new hires are expected to meet a minimum standard for AI fluency. In March 2026 it published a second version of its rubric that raised that bar.
The three traits I screen for in engineers apply here too: a systems-builder mindset, curiosity, and an understanding of how the sausage gets made. The full framework is in our guide to hiring AI talent in Charlotte. For a marketer or an operations lead, the questions change but the traits do not.
Use one scale for every question. It is the same scale as our AI engineer interview questions, so you can compare candidates across teams.
| Score | What you heard |
|---|---|
| 1 · Weak | Uses AI like a search box, or not at all. No example from their own work. |
| 2 · Partial | A real example, but the work itself did not change and nothing was checked. |
| 3 · Solid | A repeatable way of working with AI, a clear check on the output, and an honest failure. |
| 4 · Strong | Built something others on the team now use, and can say what it saved or improved. |
Zapier's published rubric also has four levels: Unacceptable, Capable, Adoptive and Transformative. Capable is its minimum for new hires. Zapier describes that level as using AI to operate at a meaningfully higher level. Its top level is about re-engineering how work happens.
Pick five or six per interview and ask every candidate the same ones. Follow up twice: ask how they know, then ask what they would change.
These four test whether the candidate changes how work gets done, or only speeds up one step.
These four test whether the candidate keeps learning without being told to.
A non-engineer does not need to explain model training. They do need a working sense of why the tools behave the way they do.
What good looks like depends on the job. These examples come from the rubrics Zapier has published for its own teams. Use them as a starting point and write your own for each role.
An interview shows how someone talks about work. A work sample shows the work. Zapier says it watches candidates use AI in real time, including how they prompt, push back on outputs and adapt. It also says a rough result with strong reasoning and real iteration is a better signal than a polished one. This is the version I would run.
Use made-up or public data for the task. Do not hand a candidate real customer records.
Fill this in right after the interview, before you compare notes with anyone.
Candidate: Role: Interviewer: Date: INTERVIEW Systems-builder mindset (Q1 to Q4) Score: _ / 4 Evidence: Curiosity (Q5 to Q8) Score: _ / 4 Evidence: How AI works (Q9 to Q12) Score: _ / 4 Evidence: WORK SAMPLE Delegation _ / 4 Description _ / 4 Discernment _ / 4 Diligence _ / 4 What they checked: What they changed by hand: Strongest signal: Biggest concern: Recommend: yes / no
The rule I use for engineers works here too: look for at least a 3 on every trait. For non-engineering roles I weigh discernment and diligence most, because those are the two that protect customers.
Building with AI Charlotte is open to every skill level, and not everyone who pitches is an engineer. At our October meetup, Bart Baldwin showed Stubbly, a link shortener he built with AI tools without a traditional coding background. If you want to meet people like that, or become one, come to a meetup. See the upcoming events, or RSVP for the December 2 meetup. More guides are on our resources page.
AI fluency is the ability to work with AI tools well. The AI Fluency Framework by Rick Dakan, Joseph Feller and Anthropic describes four parts: deciding when to use AI, describing what you want, judging the output, and taking responsibility for the result.
Ask open questions about real work, such as a task the candidate now does differently because of AI and a time a tool gave a wrong answer. Then watch a 30-minute live task. Score both on the same 1 to 4 scale and write down the evidence.
They should know which tasks to hand to AI, how to give clear context and examples, how to check output before using it, and what data must stay out of public tools. They do not need to code or know how to train a model.
That depends on the role and the company. Zapier has expected every new hire to meet a minimum AI fluency standard since May 2025. In Microsoft and LinkedIn's 2024 Work Trend Index, 66% of leaders said they would not hire someone without AI skills.
Building with AI Charlotte is a free monthly meetup at Hygge Coworking, 933 Louise Ave, open to all skill levels, with hands-on workshops and builder demos. The next meetup is Wednesday, December 2, 2026. Our guide to learning AI in Charlotte lists local classes and courses.
Last checked November 11, 2026. The questions, the 1 to 4 scale, the work sample and the hiring rule are the author's own. Spot something out of date? Tell us.
Next meetup: Wed, Dec 2 · 6:30 PM · Hygge Coworking, Charlotte. Free.