Every boardroom conversation about AI right now follows a similar script. The investment case is made. The tools are selected. The rollout is announced. And then, somewhere between pilot and production, things slow down in ways nobody planned for.
The technology is rarely the reason.
After nearly two decades in technical recruitment and workforce strategy, I have watched this pattern repeat across industries and company sizes. The organizations that struggle to scale AI are almost never struggling because the models are not good enough or the platforms are not mature. They are struggling because they do not have the people to run, govern, validate, or integrate those systems at any meaningful scale.
This is the workforce bottleneck that enterprise AI conversations consistently underweight, and it is creating real, measurable drag on returns.
The numbers behind the gap
The World Economic Forum’s Future of Jobs Report 2025 identifies skill gaps as the single biggest barrier to business transformation, cited by 63 percent of employers as a major constraint through 2030. Separately, 85 percent of those same employers plan to prioritize upskilling, yet only 32 percent express confidence that their organization already has the skills needed for long-term success.
That gap between intent and readiness is where AI ROI quietly disappears.
Nash Squared’s 2025 data shows demand for AI skills nearly doubled year-over-year, jumping from 28 percent of businesses reporting AI as a priority skill in 2024 to 51 percent in 2025. In the US alone, job postings requiring AI proficiency grew by over 1,800 percent in a recent two-year span. Supply has not come close to keeping pace. In financial services and healthcare, the two industries investing most heavily in enterprise AI, the average time-to-fill for AI-adjacent roles is running at six to seven months.
Six to seven months. For a role that, in many cases, is central to executing a technology strategy the organization has already committed to.
What “AI readiness” actually requires
Part of the problem is that AI readiness gets framed too narrowly. Companies go looking for machine learning engineers and data scientists, who are genuinely hard to find, and stop there. The skills picture is considerably wider.
Effective enterprise AI deployment requires people who can do at least six distinct things: design and maintain data pipelines that feed the models, translate business requirements into AI-appropriate problem frames, evaluate model outputs for accuracy and bias, manage compliance and auditability in regulated workflows, communicate system behavior to non-technical stakeholders, and handle the change management that comes with any workflow transformation.
Most organizations staff for one or two of these and assume the rest will sort itself out. It does not.
The Workday 2025 Global State of Skills research found that only 54 percent of business leaders say they have a clear view of the skills within their existing workforce. If you do not know what you already have, you cannot build a rational plan for what you need. And you cannot build one fast enough when a six-month vacancy is bleeding out the ROI projections you put in front of your board.
The ASEAN-specific dimension
For companies operating across Southeast Asia, this problem has additional layers.
The regional talent distribution for AI skills is highly uneven. Concentrations exist in Singapore, parts of Malaysia, and select metro areas in Indonesia, but depth drops off quickly outside those hubs. Localization requirements for ASEAN language processing add another layer of specialization that global AI vendor solutions typically do not cover out of the box. A model trained primarily on English-language data needs meaningful local adaptation to perform reliably in Thai, Bahasa, Vietnamese, or Tagalog-facing applications.
That adaptation requires people with both the technical depth to work at the model level and the local market knowledge to evaluate outputs meaningfully. That specific combination is currently one of the scarcest profiles in the region’s talent market.
Companies that are addressing this are doing so through deliberate strategies: building partnerships with universities that have strong AI programs in-country, investing in structured upskilling of domain experts who already understand the local market, and working with workforce partners who maintain pre-vetted talent pipelines rather than starting searches from scratch when a need becomes urgent.
The internal mobility answer nobody takes seriously enough
Before expanding the search externally, the most cost-effective first move is almost always a better look at what already exists internally.
SHRM’s 2025 data shows that internal talent marketplace adoption grew from 25 percent of organizations to 35 percent in a single year. That growth reflects a real shift in understanding: the person best suited to support an AI rollout is often someone who already knows the business processes the model is being applied to. A claims analyst who understands the edge cases in insurance underwriting is a more valuable AI validation resource than a generalist data scientist who has never seen the domain.
Getting that person ready for the new role requires a structured upskilling investment, not a six-month external search. The economics are straightforward. The barrier is usually organizational, specifically the assumption that internal candidates are not technical enough, without ever actually assessing them.
What the high-performing organizations are doing differently
The companies scaling AI effectively share a few operational patterns that are worth naming directly.
They define AI roles by what the work actually requires, not by credential proxies. A job description that requires a computer science degree and five years of machine learning experience will screen out a substantial portion of the people who could do the job well.
They treat talent pipeline development as part of the AI strategy, not a downstream HR task. The best time to start building the bench for a capability you need is before the urgency hits.
They work with external partners who understand AI skill requirements at a granular level, not just at the job title level. The difference between someone who has “worked with AI tools” and someone who can architect a production-ready inference pipeline is not visible on a resume without significant digging.
And they measure the right things. Time-to-fill for AI roles, quality of hire at six months, and the gap between projected and actual AI delivery timelines are all connected data points. Organizations that track them together tend to spot the talent-driven drag on AI ROI early enough to address it.
The practical bottom line
AI is not stalling because the technology is immature. In most enterprise contexts, the platforms are more capable than the organizations deploying them.
The constraint is human. It is the analyst who can bridge model output and business decision-making, the engineer who can maintain the pipeline at production scale, the compliance lead who understands what auditability means in a specific regulated context.
Finding, developing, and retaining those people is the work that determines whether an AI investment delivers what was promised. Organizations that treat that work as strategic, rather than as a procurement problem to hand off when a role opens, will pull materially ahead of those that do not.
The technology gap closed faster than most people expected. The talent gap is next. The companies that close it first will have a durable advantage that is harder to replicate than the tools themselves.
Milind Naik is Vice President of Sales and Recruiting at Compunnel Inc., with 19+ years of experience in US client recruitment. He focuses on identifying high-impact technical talent and scaling recruiting teams to meet complex hiring demands. With deep expertise in workforce augmentation and client engagement, Milind shares direct, experience-grounded perspectives on recruitment strategy, talent dynamics, and workforce solutions at Compunnel Inc.
TNGlobal INSIDER publishes contributions relevant to entrepreneurship and innovation. You may submit your own original or published contributions subject to editorial discretion.
Featured image: Campaign Creators on Unsplash
Keeping pace with industry disruption in the age of AI: 4 strategic imperatives