Evaluating AI fluency for marketing teams

Table of contents

What is AI fluency?

AI fluency goes beyond knowing what ChatGPT or Copilot can do. It means knowing when to use AI, how to direct it, and how to assess its output.

For marketing teams, AI fluency is quickly becoming a baseline expectation rather than a differentiator. According to McKinsey, demand for AI fluency in the workforce jumped nearly sevenfold between 2023 and mid-2025. Job postings now list AI fluency as a requirement for approximately seven million US workers.

As Board of Innovation puts it, AI fluency is “now a core leadership currency.” It is no longer optional for marketers or the teams they work within.

Zapier’s AI fluency rubric V2

In May 2025, Zapier open-sourced V1 of their AI fluency rubric. Hundreds of companies used it to screen candidates and develop teams. It worked — but the floor moved fast. In 2026, Zapier CEO Wade Foster released V2, raising the minimum bar and adding new dimensions.

Zapier now requires AI fluency for 100% of new hires. As Brandon Sammut, Zapier’s Chief People Officer, put it:

“To transform our business and grow Zapier’s talent density over the next two years, Zapier needs to become AI-first in everything we do.”

The three levels

Zapier defines three levels of AI fluency:

1. Capable — “I use AI to operate at a meaningfully higher level.” You use AI tools regularly and can show clear impact on quality, efficiency, or outcomes. In V2, this is the minimum bar. If someone is not meaningfully improving their work with AI, they do not meet the standard.

2. Adoptive — “I orchestrate AI and build systems that elevate how I work.” You have moved beyond one-off prompts to building repeatable AI-powered workflows. AI is embedded in how you operate daily.

3. Transformative — “I re-engineer how work happens.” You rethink entire processes, strategies, or team structures around what AI makes possible. You are not just using AI — you are redesigning work itself.

The four dimensions

Zapier evaluates these levels across four dimensions:

• Mindset — Understanding AI’s potential and adopting an experimental approach to work.
• Strategy — Identifying opportunities where AI can meaningfully improve outcomes.
• Building — Creating repeatable systems and workflows using AI, not just one-off applications.
• Accountability — Defining quality standards beforehand, critically evaluating outputs, and owning results. As Zapier frames it: “With AI, you can delegate the work, but not the accountability.”

What changed in V2

Five key shifts from V1 to V2:

1. Higher minimum standard. “Capable” no longer means simply using AI with purpose. It requires embedded, systemic use with measurable impact.
2. Trajectory over snapshot. Evaluators now assess the slope of AI adoption. Forward momentum and evidence of experimentation matter more than your current position.
3. Accountability as a core dimension. New emphasis on quality control, pre-work standards, and ownership of AI-generated outputs.
4. Manager-specific expectations. Managers must create psychological safety for AI experimentation, set clear upskilling expectations, and demonstrate workflow redesign.
5. Redesigned skills tests. Informed by Anthropic’s AI Fluency Index, tests now observe real-time AI use. Zapier values “rough results with strong reasoning and real iteration” over polished outputs without visible process.

Department-specific examples

Zapier provides role-specific criteria for every department. A few marketing-relevant examples:

• Marketing: Automating campaign brief creation and A/B test analysis with trackable engagement metrics.
• Sales/Revenue: Implementing AI-powered prospect research and email personalisation with pipeline impact data.
• Product: Applying AI to user research synthesis and feature requirement drafting with documented time savings.
• Data: Using AI for query optimisation and data documentation with performance metrics.

The full rubric across all departments is available on Zapier’s blog.

How other companies approach AI fluency

Zapier is not alone. Several organisations have published their own frameworks, each with a slightly different lens.

Salesforce: The AI Fluency Playbook

Salesforce published its AI Fluency Playbook, built from its own experience deploying AI agents as “Customer Zero” for Agentforce. The playbook organises AI fluency into three pillars:

• AI Engagement — Building employee confidence, addressing concerns, and introducing AI’s role through internal communications and change management.
• AI Activation — Embedding AI agents into existing workflows, setting usage expectations, and aligning teams on shared practices.
• AI Expertise — Developing human skills (learning agility, adaptability, communication, critical thinking) alongside agentic and business skills.

The business case is clear. Salesforce reports that employees who use AI daily see 64% higher productivity, 58% better focus, and 81% greater job satisfaction.

Captivate Talent: AI-native GTM assessment

Captivate Talent developed a framework specifically for go-to-market roles (Sales, Marketing, RevOps, and Customer Success), categorising candidates into three levels:

• AI-Curious — Experiments with tools.
• AI-Active — Implements AI in workflows.
• AI-Native — Transforms systems using AI.

Their key insight: “Most job descriptions and interview loops still focus on old skills,” while top performers blend commercial expertise with technical fluency. The framework helps hiring managers move beyond assessing outdated competencies toward evaluating genuine AI-driven operational capability.

Anthropic: AI fluency courses

Anthropic launched AI fluency courses covering framework and foundations, aimed at both individual contributors and leaders looking to build organisational capability.

Building AI fluency into your team culture

Having a rubric is one thing. Building a culture that supports AI fluency is another. Research from Forrester found that only 14% of global individual contributors score high on AIQ — highlighting a massive gap between AI availability and AI adoption.

Four practical pillars for building AI fluency culture:

1. Psychological safety. Create space for weekly team discussions about AI experiments — both successes and failures. People will not experiment if they fear looking foolish.
2. Leadership modelling. Leaders experimenting publicly gives permission for everyone else. If your CMO is sharing their prompt experiments in Slack, the team will follow.
3. Feedback loops. Regular check-ins: What is working? What is worth revisiting? What did we learn?
4. Normalised experimentation. Set clear boundaries where exploration is safe — internal drafts, sandboxed environments, low-stakes workflows.

Practical tactics that work:

• AI office hours. Open forums to share wins and learnings.
• Prompt playbooks. Shared documentation of effective prompts for real workflows.
• Monthly challenges. Small experiments with team reporting.
• Learning drops. Bite-sized insights shared via Slack or Teams.

According to IDC, organisations that foster continuous learning outperform their peers by up to 68%, with 35% revenue improvement among companies investing in learning systems.

As one practitioner put it: “Progress beats perfection. Curiosity scales.”

Why this matters for marketing teams

Marketing is one of the functions most immediately impacted by AI fluency — or the lack of it. Campaign creation, audience research, content production, analytics, and personalisation are all areas where AI-fluent marketers operate at a fundamentally different speed and quality level.

This shift is about team capability, not individual productivity. A marketing team where everyone operates at Zapier’s “Capable” level will ship faster, test more ideas, and make better decisions than a team where only one or two people know how to use AI effectively.

If you do not yet have an AI fluency framework for your marketing team, start with three steps:

1. Audit your current state. Where does each team member sit on the Capable–Adoptive–Transformative spectrum? Be honest.
2. Set a minimum bar. Decide what “good enough” looks like for your team today. Zapier’s V2 is a useful benchmark.
3. Create the conditions for growth. Psychological safety, shared learning, and regular experimentation time matter more than mandating tool adoption.

The companies building AI fluency frameworks now — Zapier, Salesforce, and others — are not doing this because it is trendy. They are doing it because the gap between AI-fluent and AI-illiterate teams is already showing up in speed, quality, and results.