Your next hire needs to know their way around AI. According to Microsoft’s latest Work Trend Index, 78% of leaders are already hunting for AI talent. But the kicker is, you need someone who can actually think with AI – not just copy-paste answers from ChatGPT.

We're here to help you spot real AI fluency in job candidates. No buzzwords, no BS. Just a practical framework to find candidates who know how to treat AI like the powerful tool that it is.

The AI reality check

Here's a stat that might surprise you: 62% of employees think AI is "significantly overhyped," according to research from GoTo.

And honestly? They've got a point.

Remember when Klarna replaced 700 workers with AI bots? Yeah, they're scrambling to hire humans again. Turns out, AI can't quite handle everything.

But we're not here to bash AI. It's incredible technology. But it also comes with some very real risks for employers. The smart move isn't overhyping or avoiding it – it's finding people who know the difference between what AI can and can't do.

🤔Still not sure if you should hire a human or outsource to AI? Don’t miss our no-nonsense guide to deciding which tasks to delegate to a real person and which ones to hand over to your AI.

A simple framework to assess AI fluency in job candidates

According to Anthropic, there are two key ingredients for effective AI fluency:

Augmentation - Employees and AI work together as “creative thinking and task execution partners.”
Agency - Employees build the AI’s knowledge and behavior patterns, rather than blindly handing it task after task.

Here’s a straightforward matrix using Anthropic’s AI Fluency Framework (a.k.a. the 4Ds) to help you identify candidates who know how to use AI beyond basic automation:

Delegation - Demonstrates clear criteria for task allocation and awareness of AI's limitations

⭐ Basic: "I use AI for writing emails"
⭐⭐ Intermediate: "I use AI for first drafts but always edit for tone and accuracy"
⭐⭐⭐ Advanced: "I delegate data analysis to AI but handle strategic decisions and client relationships myself"

Description - Shows evidence that they can iterate on prompts and provide detailed context for improved outputs

⭐ Basic: Uses simple, one-line prompts
⭐⭐ Intermediate: Adds context and examples to prompts
⭐⭐⭐ Advanced: Iterates 3-5 times, providing role context, constraints, and specific output formats

Discernment - Shares specific examples of fact-checking and quality control

⭐ Basic: "I check AI outputs before using them"
⭐⭐ Intermediate: Has a checklist for verifying facts and sources
⭐⭐⭐ Advanced: Describes catching specific AI errors and has systematic verification processes

Diligence - Shows awareness of data privacy, attribution, bias checking, or ethical guidelines

⭐ Basic: Mentions AI ethics are important
⭐⭐ Intermediate: Avoids putting sensitive data into AI tools
⭐⭐⭐ Advanced: Actively checks for bias, properly attributes AI assistance, and follows company AI policies

The goal is to find someone who treats AI like a really smart intern – helpful, but definitely needs supervision.

3 practical strategies for assessing AI fluency

Time to get practical. Here's how to spot your next AI-savvy hire in 3 straightforward steps:

1. Look for real results in the resume review

First things first. Let's decode what AI fluency actually looks like on paper. (Spoiler: It's not about dropping "prompt engineer" into their LinkedIn headline.)

What to look for:

👍Green flags

Real results: "Reduced content creation time by 40% using AI for first drafts"
Workflow wins: "Built an AI-powered customer triage system that cut response time in half"
Teaching moments: "Trained 15 team members on ethical AI use"
Efficiency gains: Specific time savings or productivity boosts
Smart integration: Shows AI as part of a bigger process, not the whole solution

👎Red flags

Buzzword bingo: "AI ninja" or "ChatGPT expert" with zero examples
Autopilot mode: Everything is "AI-generated" with no human touch
Vague claims: "Proficient in AI" (Yeah, but what does that actually mean?)
No context: Can't explain when they chose NOT to use AI
Ethics blind spot: Zero mention of data privacy or AI limitations

The best candidates don't just list AI tools – they tell stories about solving real problems. Look for the "how" and "why," not just the "what."

➡️ Find AI-savvy talent without hours of manual screening. Breezy's Candidate Match Score instantly flags the resumes that tick all your boxes.

2. Ask scenario-based interview questions

You can't spot AI fluency with generic interview questions. You need to get inside the candidate’s head and see how they actually think about AI.

These interview questions can help reveal the real AI power users in a sea of new applicants:

Describe a situation where AI could have helped but wasn't the right solution.
How would you use AI to reduce daily work interruptions?
Tell me about a time AI gave you bad advice. How did you catch it?
If AI could only do one part of your job, which would you choose and why?

Listen for candidates who talk about AI like it's a teammate, not a replacement. The ones who get it will have stories about collaboration, not automation.

Here's a quick technique from ElevenLabs talent leader Becky Lee-Roche: Give the candidate a real problem your company’s facing and ask them to think through it out loud.

"What we're looking for," she says, "is systems thinking. We want to make sure there's always an outcome and that they can tie the outcome back to the initial goal."

That's the gold right there. You want someone who sees AI as one tool in their toolkit – not the whole toolbox.

3. Give the candidate a sample task or test project

Anyone can talk a good game about AI – but can they actually use it to solve problems?

Here are 3 ways to differentiate the amateurs from the pros:

1. BS detector test

Give the candidate some AI-generated content and watch them work. Can they spot:

The made-up "facts"?
The robot-speak phrases?
The confident-but-wrong statements?

If they're not fact-checking, run.

2. Workflow design challenge

Ask the candidate to map out a process where AI and humans collaborate. You're looking for candidates who understand that AI should handle repetitive tasks, while humans focus on strategy and creativity – not the other way around.

Example: "Here's our invoice processing workflow. How would you improve it with AI?"

👍Good answer: "AI handles data entry, humans verify exceptions and build relationships."

👎Bad answer: "Just automate everything!"

3. Prompt evolution exercise

Give the candidate a basic prompt and watch them refine it.

Start with: "Write an email"

Watch them turn it into: "As a customer success manager, write a friendly follow-up email to a frustrated client who experienced a service outage. Acknowledge their concerns and offer concrete next steps."

Here are some examples of projects to test AI skills for specific role requirements:

Marketing

Challenge: Create an AI-assisted campaign plan for a product launch
What to watch for: Do they use AI for research and ideation while maintaining brand voice? Can they spot when AI suggestions don't align with your brand or target audience?

Customer Service

Challenge: Triage a set of AI chatbot escalations
What to watch for: Can they identify patterns in what the AI missed? Do they suggest improvements to prevent future escalations?

Operations

Challenge: Design a hybrid AI-human workflow for invoice processing
What to watch for: Do they identify which steps need human oversight? Can they build in quality checks without creating bottlenecks?

Sales

Challenge: Use AI to personalize 10 outreach emails without sounding generic
What to watch for: Do they maintain an authentic voice while scaling personalization? Can they spot AI suggestions that would damage relationships?

Recruit AI-ready candidates

AI is great at a lot of things, but it's terrible at being human. It can't read the room. It doesn't get nuance. And it definitely can't handle that customer who's having a meltdown because their order is late.

You need someone with killer soft skills who knows how to make AI work for them — not the other way around.

Ready to put these steps into action? Breezy's got your back. Our AI-powered screening tools help you find the humans who know how to work with AI – meta, right?

Screen smarter, hire faster, and build a team that's ready for whatever comes next. Start your free 14-day trial and see the difference.