Fixing the Talent Gap in AI: A Skills-First Strategy for Workforce Planning

Vishnu Shankar, VP of Data and Research at Draup, offers a “skills-first” strategy for workforce planning that can help companies address the AI talent gap. This article originally appeared in Insight Jam, an enterprise IT community that enables human conversation on AI.

The demand for AI skills has surged faster than workforce planning leaders expected. Salaries have spiked 56 percent in North America, 52 percent in the UK, and 35 percent in India, and advanced AI roles now require premiums of up to 47 percent above standard technical positions. Yet, the bigger issue is attrition. External hires cost 25 percent to 30 percent more than internal promotions and are half as likely to stay beyond 18 months, pushing companies back into an increasingly overheated talent market.

External hiring simply can’t keep up.

• AI skill demand in software jobs has surged 50 percent in two years (Draup research), making external supply far too thin to rely on.
• Core skills like SQL, Python, and Java remain heavily contested, limiting the pool for AI-adjacent hiring.
• AI tools and coding copilots shift work but have not significantly reduced headcount needs, keeping demand for skilled engineers high.
• Nearly 3 million technical roles will open by 2033 due to retirements, shrinking the external pipeline even further.

Talent stakeholders now face a choice: continue absorbing the disadvantages of an overheated external market, or build internal pipelines that create stability, increase productivity, and reduce long-term costs.

Reskilling is the Answer, and the ROI Backs It 

Draup’s research shows that reskilling an existing employee is significantly more cost-effective than hiring externally, with total costs estimated to be 23 percent lower. The advantage stems from multiple, mutually reinforcing levers.

• Lower cost: Internal development avoids external hiring premiums and reduces total spend.
• Higher retention: Employees who grow internally stay longer, lowering churn and replacement costs.
• Faster productivity: Internal talent reaches full contribution much sooner.
• Compounding value: Skill investments remain inside the company instead of walking out the door.
• Strategic flexibility: Reskilled teams adapt as AI roles evolve, reducing reliance on volatile labor markets.

But to reskill at scale, organizations need a precise understanding of the skills they have, the skills they need, and how those skills are evolving. This is where clarity of skills becomes essential.

Draup’s analysis of talent data across Fortune 100 companies highlights why precision matters. While 65 percent of software and mathematics talent already possesses foundational AI capabilities such as programming and libraries, more than one in four already demonstrates capability across at least two of the four core AI capability layers spanning development, deployment, and application. Nearly one in ten already spans three or more capability layers of the AI stack, a level of breadth typically associated with deployable AI roles. The challenge is not a lack of skills, but a lack of visibility into how close employees already are to AI-critical roles.

Skills Clarity: The Foundation of Scalable, Future-Ready Workforce Planning 

Skills clarity is the core enabler of internal talent development. With an accurate view of existing AI capabilities, emerging gaps, and adjacent skill pathways, workforce teams can build reliable internal pipelines and target reskilling where it has the greatest impact. But as AI accelerates the convergence of cloud, data, machine learning, and security, skill requirements are evolving too quickly for static frameworks to keep pace. Draup’s longitudinal analysis shows that across the top 100 software and mathematics roles, roughly 25 percent of critical skills change each year, with peak periods seeing as much as a 33 percent year-over-year shift.

This is why organizations must move from fixed competency models, which age out within a year, to dynamic skills architectures that update continuously and show who can be developed, for which roles, and on what timeline. For example, across roles such as cloud engineers, full-stack engineers, and solutions architects, about one-third of the workforce already spans multiple AI capability layers, positioning them well for ML Ops, AI application engineering, and applied AI roles with targeted reskilling.

This need for precision is transforming Strategic Workforce Planning (SWP). Traditional dashboards cannot keep up with the pace of skill evolution. The new SWP stack relies on intelligence-driven components that shift planning from backward-looking reporting to forward-looking decision-making.

Skills clarity enables four essential capabilities for internal talent development:

• Dynamic Skills Architecture: Real-time skill definitions that identify employees closest to emerging AI, cloud, data, and security roles, such as cloud engineers already meeting most MLOps requirements.
• Workload & Task Redesign: Breaks work into tasks and workloads instead of static job titles, enabling more accurate automation modeling, cost forecasting, and identification of employees who can be reskilled, given that 40-50 percent of workloads in AI-adjacent roles already overlap with existing cloud, data, and software engineering workloads (Draup analysis).
• Multi-Agent Control Planes (MCP): Intelligence that surfaces the highest-impact reskilling opportunities and flags where internal development outperforms external hiring.
• Copilots & Enterprise Workflows: AI-enriched work environments that support continuous learning and faster skill acquisition.

With skills clarity in place, internal talent development becomes scalable, predictable, and more cost-effective than external hiring.

Next Course of Action for Strategic Workforce Planning  

Enterprise workforce planning teams that treat talent development as a strategic capability rather than an HR initiative consistently outperform others on retention, productivity, and cost. Most companies already have employees who are closer to AI readiness than they realize. The path forward is focused and practical:

1. Use skills, intelligence, and global labor market data to inventory current skills and map employees to AI-adjacent roles. Draup’s analysis shows that many are only one or two steps away once existing tasks and abilities are mapped to AI-adjacent roles rather than inferred solely from job titles.
2. Build clear progression pathways such as SQL → Python → LLM pipelines or reporting → predictive analytics, and use AI-powered talent intelligence platforms to benchmark and track progress. In practice, this allows workforce teams to stop enrolling entire job families into generic AI programs and instead target a small set of missing capabilities for specific role transitions.
3. Make internal mobility visible and structured by posting opportunities, setting manager expectations, and offering sandbox environments for experimentation. We’ve seen organizations accelerate internal transitions when employees can apply new skills through short-term projects before formal role moves.
4. Use external hiring strategically, bringing in specialists when required to seed new capabilities and mentor internal teams rather than to replace core talent needs. This shifts hiring from volume-driven replacement to targeted capability acceleration.

Seventy-seven percent of employers plan to prioritize reskilling by 2030. The advantage will go to those who act early. We have seen that the real divide is not between perfect and imperfect workforce plans, but between organizations that treat talent development as a long-term strategic asset and those that treat it as administrative work.