The 7% Problem: Why Companies Are Bleeding Money on AI While Ignoring Their People

Every executive boardroom in 2025 has the same refrain: "We need an AI strategy." Budgets are approved at record pace. Vendors are signed. Models are trained. GPUs are humming in data centers around the world. And yet, according to multiple industry analyses, only 7% of total AI investment is directed at the human factor — the people who design, deploy, govern, and ultimately live with these systems every day.

That is not a rounding error. It is a strategic crisis hiding in plain sight.

In this post, we'll unpack why this imbalance exists, what it is actually costing organizations, and how forward-thinking leaders are rebuilding their AI investment portfolios to put people back where they belong: at the center of the strategy.

The AI Investment Paradox

The numbers are striking. Global enterprise AI spending is projected to surpass $300 billion by 2027, according to IDC's Worldwide AI Spending Guide. Yet survey after survey — from McKinsey, BCG, Deloitte, and the IBM Institute for Business Value — consistently shows that a sliver of that spend is dedicated to workforce readiness, training, change management, and organizational design.

When we say "the human factor," we mean the full stack of people-related investments:

• Reskilling and upskilling programs for existing employees
• Hiring and retention of AI-literate talent
• Change management to redesign workflows around AI-augmented processes
• Governance, ethics, and compliance training
• Cross-functional collaboration between technical and business teams
• Culture and leadership development to shepherd AI transformation

All of this combined? Roughly 7 cents on the dollar.

Where the Other 93% Goes

To be fair, the 93% isn't being wasted — but it is being heavily skewed. Typical AI budget allocations look something like this:

• 40–50% — Infrastructure: GPUs, cloud compute, data platforms, MLOps tooling
• 20–25% — Software: model licenses, SaaS AI products, integration platforms
• 15–20% — Data engineering: pipelines, labeling, warehousing, quality
• 8–10% — Talent acquisition (technical hires only)
• 5–8% — Pilots, proofs of concept, vendor evaluation
• 3–5% — Risk, compliance, and governance tooling
• 2–3% — Training and enablement
• 1–2% — Change management and culture

Notice how "training" and "change management" — the two categories most likely to determine whether AI actually delivers ROI — sit at the bottom of the stack. The infrastructure is gleaming, the models are powerful, and the people expected to use them are largely unprepared.

The Hidden Costs of Neglecting People

Why does this imbalance matter? Because AI doesn't fail at the model layer — it fails at the adoption layer.

1. The 70% Failure Rate

McKinsey's research has repeatedly shown that roughly 70% of digital transformations fail, and AI initiatives follow the same pattern. The top-cited reason is not technical debt or bad models. It is people and process issues: lack of buy-in, insufficient skills, unclear ownership, and cultural resistance.

A $5 million model that nobody on the operations team trusts to use is worth exactly $0.

2. Shadow AI and Governance Gaps

When companies don't invest in enablement, they don't get fewer AI use cases — they get unmanaged ones. Employees quietly adopt consumer AI tools, paste sensitive data into public chatbots, and build shadow pipelines. The result is the worst possible outcome: AI is being used everywhere, but the organization has no visibility, no control, and no ability to capture the value.

3. Talent Attrition

AI-literate professionals are among the most mobile workers in the global economy. A 2024 LinkedIn report found that AI-skilled employees are 3.2x more likely to change jobs than the average worker. Companies that invest in infrastructure without investing in their people are essentially building beautiful factories with no workers — and watching their best employees walk out the door to competitors who actually fund their development.

4. Productivity Theater

We have entered the era of productivity theater: employees are adding AI tools to their workflow, but most are using them at a superficial level. A recent Harvard Business School study found that consultants using AI without proper training saw only marginal productivity gains, while those with structured AI training improved performance by 40% or more. The same tools. Wildly different outcomes. The variable was the human, not the model.

A Framework for Human-Centered AI Investment

So what should the right allocation look like? At Tanok Tech, when we work with clients on AI strategy, we recommend a framework we call the "40-30-20-10 Rule":

Allocation	Focus	Example Investments
40%	Technology & Infrastructure	Compute, data platforms, MLOps, vendor licenses
30%	Talent & Skills	Reskilling programs, AI literacy for all roles, hiring AI translators
20%	Process & Workflow Redesign	Change management, new operating models, success metrics
10%	Governance, Ethics & Culture	Responsible AI policies, risk frameworks, leadership coaching

This doesn't mean infrastructure gets short-changed — 40% is still the largest single category. But it elevates the human and process categories from afterthoughts to strategic pillars.

The Three Roles Every AI-Ready Organization Needs

Most companies invest in two of the following three roles and wonder why their AI initiatives stall. You need all three:

1. AI Builders

These are your data scientists, ML engineers, and AI architects. They build and maintain the models. You cannot skip this investment — but you also cannot win with it alone.

2. AI Translators

This is the role most companies forget. AI translators are domain experts (marketers, supply chain managers, financial analysts) who deeply understand AI capabilities and the business. They bridge the gap between technical teams and operational leaders. Companies like Deloitte and PwC have built entire practices around this role for good reason — it's often the difference between a model that sits in a notebook and one that transforms a P&L.

3. AI Literates

Every employee needs a baseline understanding of what AI can and cannot do. Not everyone needs to code. But everyone should understand how to prompt, how to evaluate AI output, how to recognize bias, and when to trust vs. override a model.

Practical Steps to Rebalance Your AI Budget

If your current allocation looks like the 7% problem, here's a six-month plan to shift course without abandoning your technical roadmap.

Month 1–2: Audit and Diagnose

• Survey your workforce on AI confidence and current usage
• Map AI use cases against the skills required to execute them
• Identify the top three workflow gaps where AI is being underutilized

Month 3–4: Launch an AI Literacy Program

• Start with a mandatory foundational course for all employees (4–6 hours)
• Build role-based learning paths for high-impact teams (sales, ops, finance, engineering)
• Identify and train AI champions in each business unit — internal evangelists who can coach peers

Month 5–6: Redesign Two Pilot Workflows

• Pick two high-value processes and redesign them end-to-end with AI at the core
• Measure both technical metrics and human metrics (adoption rate, time saved, satisfaction)
• Document and socialize the wins — social proof is the most powerful change management tool you have

Real-World Lessons

A few patterns from our client work are worth sharing:

> A mid-market insurer came to us with a $4M model risk platform. After deployment, adoption hovered around 12%. We diagnosed the issue in two weeks: the underwriting team had zero training on how to interpret model outputs, and the model's explanations were technical jargon. The fix wasn't a better model — it was a six-week enablement program and a redesigned UI with plain-language explanations. Adoption jumped to 78% within four months.

> A global retailer deployed a recommendation engine across 14 markets. Eight markets saw immediate lift. Six saw flat or declining performance. The difference? The eight successful markets had localized training and AI champions in each region; the six underperformers had been told to "just use the tool." Same technology. Massive outcome gap.

> A B2B SaaS company invested heavily in AI features but saw churn spike. Their customer success team couldn't explain the AI features to customers, and customers churned out of confusion. After embedding "AI fluency" as a core CS competency and training the team, churn returned to baseline within a quarter.

The pattern is unmistakable. Technology is necessary. Technology alone is insufficient.

What Smart Investment Looks Like in 2026

As we look ahead, the companies pulling ahead are not necessarily those spending the most on AI infrastructure. They are the ones asking a fundamentally different question: "What is the total investment required to make AI work in our organization — including the human system it operates within?"

That means tracking metrics that CFOs rarely see today:

• AI adoption depth: not just who has access, but who is using AI meaningfully in their daily workflow
• Time-to-competence: how long until a new hire becomes productive with your AI stack
• Workflow redesign coverage: what percentage of AI-impacted processes have been formally redesigned
• Internal mobility into AI roles: are you growing talent or just hiring it?

These metrics force the conversation away from "what did we build" toward "what is working in our organization" — and they naturally rebalance budgets toward people.

Conclusion: The 7% Problem Is Solvable

The 7% statistic is not destiny. It is the result of choices — and choices can be reversed.

Every dollar shifted from duplicative infrastructure spend to enablement, training, and change management is, in our experience, a dollar that returns 3–5x in realized AI value. The math is simple: an AI strategy that the organization cannot operate is not a strategy at all.

If you are leading AI transformation in your organization, the most important question you can ask this quarter is not "What should we build next?" It is "Who is ready to use it — and what have we invested to make them ready?"

---

Ready to rebalance your AI investment strategy? At Tanok Tech, we help organizations design AI programs that work in the real world — where humans and algorithms have to coexist productively. [Talk to our AI strategy team](#) and let's map your human-AI readiness together.