There's a growing gap in many organizations where, on one side, some teams use AI tools like AI assistants, AI-powered workflows, or agents that can handle multi-step tasks autonomously — and, on the other side, there are employees who aren't sure how to use these tools well or why it matters that they do. This is the gap of AI fluency.
Airtable research found that 56% or organizations are still in the earliest stages of human-agent collaboration, where most use is at the individual level instead of supporting the wider organization. Closing these gaps becomes more important as leaders look at how their teams will run most effectively as work continues to change.
What is AI fluency?
AI fluency means having the ability to work with AI tools and systems — understanding their potential and limitations, and how to apply them to practical workflows. WhileAI fluency doesn’t require non-technical teams to build models from scratch or to write code, they should understand how to use AI intentionally to produce the best results. That begins with basic AI literacy,, and then builds on that knowledge to a level where you can work with AI tools or agents, evaluate outputs, and adapt as the technology evolves.
Today, according to Anthropic’s Economic Index, “the most common expression of AI fluency is augmentative—treating AI as a thought partner, rather than delegating work entirely.” Yet this is expected to change over time, and it’s a skill that takes time to develop.
What AI fluency looks like in the workplace
Let’s take a look at what AI fluency looks like, particularly for non-technical teams.
- Knowing when to apply AI (and when not to): When starting to delegate tasks to AI, the best candidates are often repetitive, knowledge-heavy, multi-step workflows, whereas tasks that require human judgment should have built-in approval steps and guardrails to ensure security, accuracy, and usability.
- Writing effective prompts: Getting useful output from an AI tool takes practice. Your teams need to provide agents with the context, memory, and skills from a shared system of record in order to create accurate outputs. AIneeds clear and specific instructions about format, and what it can reference or should avoid, followed by iteration. Consider that a junior employee doesn’t learn if you take their work and overwrite it; instead, you have to coach them on what to do better next time.
- Evaluating AI output with a critical eye: AI can hallucinate, oversimplify, and miss nuance that’s important to your business, especially when it doesn’t have memory or skills based on a shared system of record. Part of this goes back to the prompting and guardrails in place: what it can access, whether it has too much or too little information to pull from, poor data quality, and so on. But fluent users know this and don’t treat AI output as fact. Instead, they verify, edit, and apply their own expertise before treating the output as final.
- Understanding AI limitations: Many solutions, including Airtable, offer low- or no-code tools to make it easy to work with AI and build useful workflows. Still, it’s important to learn how your solution works under the hood. AI tools have context windows that key knowledge can sit outside of, for example. They can also reflect biases in their training data or gaps where AI makes guesses instead of relying on facts. This awareness shapes how AI is applied to your workflows.
- Staying current: AI evolves fast, so part of fluency is adopting a mindset and actual practices around continuous learning. This means building in regular audits to ensure your AI tools are accessing the most up-to-date information, but also experimenting and encouraging teams to work with new tools over time.
4 ways leaders can develop AI fluency within their org
For leaders looking to support AI fluency within their organization, here are five steps you can take:
1. Make AI learning a priority, and part of the day
With AI adoption primarily at the individual level, people are generally figuring it out on their own. According to Airtable research, 44% of orgs get stuck at the “Assist” stage, where individuals see productivity gains but still step in to make decisions and execute. They don’t quite graduate to the next level where they’re adapting AI to delegate workflows that benefit the entire org. Leaders must treat AI skill-building the way they'd treat any other critical capability: with dedicated time, structured resources, and clear expectations.
2. Define clear expectations that map to specific roles
AI fluency looks different across roles. A marketer's fluency needs are different from a product manager's and certainly from an engineer’s needs. For each team, leaders need to be clear on expectations, benchmarks, and goals — and provide training and resources to help teams succeed. Consider the following:
- What does an AI-fluent person in this function actually do?
- What tools do they use?
- What decisions are they making differently?
For example, you may need your marketers to be very good at prompt writing for campaign analysis vs. knowing how to vibe code a sophisticated application. Teams need leaders to help them see their roles within a new light so that they aren’t guessing at how to spend their time. Try mapping this for a few high-impact roles.
3. Assess readiness and allow room for experimentation
AI trust and confidence comes from learning from real-world applications. Leaders must begin by assessing their AI readiness, to understand what data preparation and infrastructure needs to be put into place. Beyond this, teams need dedicated time to explore AI tools without pressure to show immediate productivity gains, as well as a collaborative space to share what’s working or where they learned an important lesson.
4. Lead by example
Leaders visibly using AI — experimenting with it, talking about what they're learning, being honest about where it falls short — signal to their teams that it’s a priority that they’re invested in. Teams feel pressure to deliver, but it helps when leaders step in and show clear intent and understanding around what they’re up against, helping to remove roadblocks. The good news is that many vendors you already use may provide resources to help your teams scale up their AI usage.
Building an AI-fluent organization, using Airtable
The business landscape is moving toward a reality where AI fluency must be the norm. AI delivers powerful results when applied strategically, and Airtable's AI workflow platform is built to give teams scalable, impactful results with ease of implementation, thanks to no-code and low-code tools and shared workspaces where humans and AI get work done. Your team can also increase their AI fluency with free resources like the Airtable Academy AI App Builder certification for a structured path to build AI skills within Airtable.
Build AI fluency with Airtable Academy
Frequently asked questions
AI literacy means understanding what AI is at a fundamental level. This can include researching different models and exploring basic functionality. AI fluency means that someone has invested time into understanding how to use specific AI tools effectively in their day-to-day work.
The AI Fluency Framework defines four interconnected core competencies: Delegation, Description, Discernment, and Diligence. Delegation is deciding what to give AI; Description is how you clearly communicate with it; Discernment is evaluating what it produces; and Diligence is taking responsibility for the outcome. Together, these competencies describe what it looks like to work with AI in a way that's effective, efficient, and ethical.
There’s more than one framework to define AI fluency, but most include a few consistent elements. These are: understanding AI capabilities and limitations, the ability to prompt AI tools effectively, critical evaluation of AI-generated output, and a commitment to continuous learning as technology evolves. Applied to the workplace, it also includes knowing how to integrate AI into actual workflows and when human judgment should take precedence.
Start by assessing your org's AI readiness — this means evaluating your data preparation, infrastructure, and where teams currently sit in their AI adoption journey. From there, define role-specific benchmarks: what does an AI-fluent marketer actually do differently from an AI-fluent product manager? Progress can be tracked by how well teams move from individual-level AI use toward delegating and adapting AI for org-wide workflows.
Several factors stall AI pilots. Teams may lack the data infrastructure needed to support reliable outputs. There may also be a confidence gap — it can feel risky to apply AI to real workflows before trust is established. Organizations that create dedicated time for low-stakes experimentation, paired with a collaborative space to share learnings, are better positioned to move pilots into production successfully.
