← Wiki

The Proficiency Gap

The gap defined

Brotman and Sack identify a structural mismatch: AI capabilities are advancing exponentially while organizational adoption remains linear. The result is a proficiency gap — the distance between what AI can do and what most leaders know it can do — that widens with each model generation.

The gap manifests at two levels: At the individual level, 80% of leaders Brotman and Sack spoke with either didn’t use ChatGPT regularly or used only the free version of ChatGPT-3.5. They used it occasionally to write an email or job description. They “all wanted to learn more about how to better incorporate gen AI into their workflow, but didn’t know where to start.”

At the organizational level, Deloitte’s 2023 AI Institute survey of 2,800 C-suite executives found only 1 in 5 believe their organization is “highly prepared” to address AI skill needs. BCG’s concurrent survey of 1,400 C-suite executives found 90% were “observers” — experimenting only in small ways, “ambivalent or dissatisfied” with their progress.

The middle era

Brotman and Sack identify a coming “middle era” between current AI (GPT-4 class) and full AGI — a period when models become dramatically more capable but organizations haven’t yet restructured around them. Mustafa Suleyman frames this as “ACI” (Artificial Competent Intelligence): models so good they can competently handle almost any task, enabled by 100× more compute.

The middle era creates a code-red urgency for closing the proficiency gap. Leaders should move from yellow to red alert because the gap will widen faster than most expect.

Literacy → Proficiency → Fluency

Both individuals and organizations progress through three stages of AI capability:

StageIndividualOrganization
LiteracyUses AI for basic search, simple emailsBasic content creation, chatbots, cost cutting
ProficiencyCustom GPTs, personal projects, complex tasksAdvanced content, cross-department integration, workflow automation
FluencyDeep understanding of capabilities and limits, novel solutionsStrategic decision-making, margin improvement, new business offerings, companywide mindset

Most organizations are stuck between literacy and proficiency. The gap compounds: organizations at fluency learn faster and extend their lead over those still at literacy.

The blank chat box problem

A counterintuitive source of the gap: the simplicity of the interface. The chat box is “hauntingly simple — just a blank input chat box, with a blinking cursor, staring at you.” This simplicity creates a paradox:

  • No learning curve to start using it
  • Enormous hidden capability that most users never discover
  • No user manual for “the most powerful software ever developed by humans”

Most leaders treat the AI chat box like a Google search box — typing simple queries when they could be uploading documents, analyzing spreadsheets, conducting multi-step strategic analyses, or creating custom AI applications.

Why the gap matters

The proficiency gap creates compounding competitive disadvantage. Gates: “The risk for businesses that delay embracing AI is not merely falling behind; it’s becoming obsolete.” The Harvard/BCG study demonstrates what closing the gap yields: up to 25% productivity gains and 40% quality improvements for tasks within AI’s capability frontier.

An associate at a top US bank completes his week’s work in a day and a half using ChatGPT, spending the rest studying AI. His superiors are “pleased with his performance yet remain unaware of his AI hack.” This individual proficiency gap — some employees 5× more productive, invisible to management — creates a hidden organizational risk.

The CEO proficiency split

Levie observes a bifurcation among public company CEOs: a cohort of founder-led or technically-oriented leaders who are “clearly wired in, tapped in” — spending weekends in Slack channels and WhatsApp groups “working with Claude Code or Codex building stuff” — and a cohort that acknowledges, as one peer told Levie, “we don’t have the AI chops in house. We might need to bring it in.” The proficiency gap is not uniform even at the top of the hierarchy.

The speed of change compounds the gap. Levie describes a cadence “on a multiple times a week cycle” where staying current requires following “the practitioner who’s the engineer at the agent sandbox company and their long form article on how they are handling memory and the harness” — information sources that are unrecognizable to leaders trained on quarterly analyst reports and McKinsey decks. Leaders who cannot track at this speed fall behind in ways that their organizations cannot compensate for.

Closing the gap

The AI Transformation Playbook provides the structured approach to closing the gap. Principles from Brotman and Sack:

  1. Education first — AI literacy is the prerequisite for everything else
  2. Learning by doing — prompt contests, hackathons, and daily use accelerate proficiency faster than lectures
  3. Show don’t tell — working prototypes convert skeptics faster than strategy decks
  4. Make it cultural, not technical — frame the transformation as mindset shift, not IT project