Workera, the trusted skills data layer for the enterprise, published the 2026 AI Skills Enterprise Benchmark Report. The report represents more than 88,000 assessments from some of the world’s most advanced enterprises and the United States federal government across 14 AI and data capabilities. According to the data, only 13% of employees are Accomplished in Agentic AI skills before any upskilling, the lowest benchmark across all 14 capabilities measured. Upskilling works: 81% of employees become Accomplished in Responsible AI after targeted upskilling, though only 25% start there.
“Talent is still the greatest multiplier in the AI era, that’s why OpenAI and Anthropic are hiring at such a pace.” said Kian Katanforoosh, CEO of Workera. “But most companies can’t afford to attract top AI talent. What they can do is act now to make their workforce AI-ready, and that requires more than standard training. It requires rigorous analysis of capability gaps, accurate measurement, and coaching systems that help people improve. Our benchmarks reflect what individuals from leading organizations actually know and can do, not what courses they completed or what they said about themselves. My hope is that this report gives leaders a clearer view of the skills that matter now and the human potential they need to build next.”
Assessments were conducted using Elo, Workera’s AI-native, multimodal skills assessment platform that generates job-relevant scenarios, delivers real-world assessment experiences, and scores responses using rubric-anchored AI evaluation, overseen by human assessment scientists at every stage. Workera assessments are designed, deployed, and evaluated by a team of AI scientists led by Taylor Sullivan, PhD, and Marian Pitel-Wali, PhD.
Across 14 capabilities, the data reveals a consistent pattern: workforces are strongest where familiar skills apply to new tools, and most underprepared where the technology itself is newest. Findings from the 2026 AI Skills Enterprise Benchmark Report include:
- Enterprises are strongest where soft skills meet new technology. The highest enterprise benchmarks were Data Storytelling Essentials (231 on a 300-point scale), AI and Data Communication (230), and Responsible AI Essentials (229). These capabilities have a lower technical barrier to entry, as employees can apply communication, ethics, and reasoning skills they already have.
- The technical gap is a strategic risk. The two weakest benchmarks were Deep Learning Fundamentals (163) and Beyond LLMs: Prompts, Agents, and Retrieval-Augmented Generation (RAG) (185). Only 13% of employees are Accomplished in the latter before any upskilling. Organizations that leave this gap unaddressed will concentrate advanced AI capability in a handful of people, creating bottlenecks, increasing project risk, and falling behind competitors who are closing it faster.
- Upskilling works, but only if you measure first. Across every capability in this report, targeted upskilling produced dramatic results. Employees who upskilled in Data Visualization and Storytelling improved by 72% on average. AI Essentials: 47%. Generative AI Essentials: 53%. The pattern is consistent: the gap between first score and current score is large, and it closes quickly with precision development. The organizations realizing this ROI started by measuring.
- Data skills outperform AI skills, for now. Both enterprise and best-in-class benchmarks are slightly stronger in Data Fundamentals than AI Fundamentals. Enterprise employees have been working with data for years, while AI is newer terrain. This gap will narrow as AI tools become embedded in daily work, but it underscores why enterprises that invested in data literacy early have a head start on AI readiness.
The benchmarks are particularly useful because despite 85% of L&D leaders saying they’re confident in self-reported skills data, the evidence tells a different story. Only 11% of employees can accurately assess their own skill levels, Workera’s research shows, and nearly 7 in 10 either overestimate or underestimate what they can do, posing a risk to fair promotions, targeted learning, and AI readiness. Workera aims to deliver a precise X-ray of a workforce’s skill gaps so enterprise executives can develop an upskilling plan with confidence.
“Accurate skills verification is the essential foundation for today’s talent and upskilling strategies.” said Bernardo Fonseca Nunes, PhD, Data & AI Transformation Specialist at Workera. “Skills metrics not only provide a precise starting point for each employee’s talent roadmap; they allow leaders to benchmark their organisation against industry peers and build a clear line of sight into project success based on skills. In the AI age, enterprises win or lose based on their learning velocity. More important than what your teams know today is how fast they can close tomorrow’s skills gaps.”
Research Methodology
The report is based on 88,753 assessments from 32,422 individuals, representing some of the world’s leading enterprises. The data spans professional services, pharmaceutical, medical, financial services, and consumer packaged goods industries, as well as the U.S. federal government. Workera assesses users on a 300-point scale, from Beginning (1–100) to Developing (101–200) to Accomplished (201–300).
The 2026 AI Skills Enterprise Benchmark Report is available today.
Explore Hrtech Articles for the latest Tech Trends in Human Resources Technology
