Why AI Transformation Needs a New Kind of Leadership, Not Just New Technology

The biggest barrier to AI transformation isn't technology, it's leadership and culture. 91% of data leaders at large companies say the barrier is culture, only 9% say technology. That means buying a stronger AI model doesn't touch the real problem. What needs to change is who makes decisions, who's accountable, and how work gets redesigned.

Disclaimer up front: this is my own read on a few surveys I've been through lately, take it as one operator's view, not gospel.

I run a small company I've bootstrapped since 2019, wearing both the operator hat and the owner hat, no formal management pedigree to speak of. I've been reading a lot of AI transformation reports the past few months, and one number stopped me: 91% of data leaders say the barrier is culture, only 9% say technology. This post is my case for why that number means a lot of companies are buying the wrong thing, and what a business leader should do differently. One note before we start: this is my own view, not a universal truth.

What's the real barrier to AI transformation?

The biggest barrier to AI transformation is organizational culture, not technology. In MIT Sloan's survey, 91% of data leaders at large companies named culture and change management as the primary barrier to becoming a data-driven organization. Only 9% named technology (MIT Sloan Management Review, 2026).

That number flips how most companies are actually spending. Money flows into models, infrastructure, licenses. Very little flows into changing how people work.

And here's the part worth sitting with. If 91% of the problem sits in people and culture, buying stronger technology is only ever addressing that other 9%.

Why doesn't buying more technology solve it?

AI technology is now something anyone can buy. Every company can get the same model from the same handful of vendors. Competitive advantage no longer lives in which tool you have.

BCG names this ratio directly in its 10-20-70 framework: roughly 10% of a successful AI transformation's effort sits in the algorithm, 20% in data and technology, and 70% in people, process, and organizational change (BCG, 2026).

Two independent sources point at the same thing. MIT Sloan measures the barrier. BCG measures where value gets created. Both say most of it sits outside the technology itself.

So what can't be bought? The organizational capability to put AI into your company's actual processes:

• Who's allowed to act on an AI output, and who still has to pick up the phone to double-check it.
• Which processes you redesign around AI, versus which ones just get AI pasted on top of the old way of working.
• Who's accountable when AI gets it wrong, instead of blaming "the system."

What does the new kind of leadership AI requires actually look like?

The new kind of leader is someone with the authority to redesign how work gets done, not someone who just signs off on the software purchase. Most CIOs today don't have the time or the mandate to own the culture-and-organization side, and that's exactly the part that blocks AI success most often.

MIT Sloan proposes a new role for this gap: the chief innovation and transformation officer. This is a leadership role with the authority to both redesign processes and manage cultural change, not just sign off on technology budgets the way a traditional CIO does. It owns exactly the space the traditional technology role leaves empty (MIT Sloan Management Review, 2026).

Your company doesn't necessarily need a brand-new C-level title. What matters is that someone actually holds three real forms of authority:

1. Authority to redesign a process, not just propose it.
2. Authority to change how performance is measured and rewarded, so the incentives actually favor using AI.
3. Authority to decide who's accountable for an AI outcome.

Give someone the budget without these three, and they'll buy tools. They won't change how the company runs.

Take my own recruiting company, 24staff.net. We started as a tech company, so we've always had the mindset of building AI into daily workflows early. In candidate sourcing, one task repeats at scale: sourcing candidates and matching them against the profile in a client's job description.

Candidates used to default to messaging a headhunter directly on Zalo. The headhunter would download the CV and eyeball it manually. Beyond building an AI-driven applicant tracking system, we had to change that workflow itself.

Now we send candidates a link to upload their CV. Their CV or portfolio flows straight into the system, where AI reads it, analyzes it, and scores how well it matches the client's job description. That process used to take roughly 5 minutes per candidate, and with 100 CVs against one job description, the workload was brutal. Now it takes about 3-5 seconds for AI to produce a match score, and the headhunter's attention goes to the highest-scoring candidates only.

Some of our clients have asked us to integrate AI directly into Zalo chat. That's not actually technically feasible yet, though, since Zalo hasn't opened an API for personal Zalo chats.

Where does Vietnam stand in this picture?

Vietnam's workforce is more ready for AI than most of the world. Its organizations haven't caught up. Vietnam leads ASEAN with 39% of knowledge workers as proficient AI users, more than double the 16% global average (Microsoft Work Trend Index, June 2026).

It's good news and a warning at once. Workers are moving fast. Leadership and governance models are moving a lot slower.

That gap is the same culture barrier MIT Sloan measured, just seen from Vietnam's side. Employees use AI to move faster, but managers still pick up the phone to call a department head and double-check the number (Thanh Nien, 2026).

When leadership doesn't trust the number, every AI investment behind it turns into waste. The problem isn't that the model isn't good enough. The problem is that decisions still run on the old track.

What should a business leader actually do differently?

If 91% of the barrier sits in culture and leadership, the first thing to change is where you put your attention, not your technology budget. Here are three concrete things I think a decision-maker should do differently:

1. Flip the attention ratio. If you're spending 90% of your time talking about tools, shift most of that time to process, decision rights, and incentive design instead.
2. Pick one person to hold all three forms of real authority. Not the person who approves the budget, but the person allowed to redesign how work gets done and reassign accountability.
3. Start with one process, don't try to swallow the whole company. Pick a process you actually control, redesign it around AI from the ground up, measure the real result, then scale it.

Picking a tool is the easy part. Changing how people work and decide is the hard part. And that's the part that actually decides whether your AI transformation succeeds or sinks.

Frequently asked questions

What's the biggest barrier to AI transformation?

Culture and leadership, not technology. MIT Sloan's survey found 91% of data leaders at large companies say the barrier is culture and change management, only 9% say technology.

Why doesn't buying more AI technology fix the problem?

Because the tools are now available to anyone from the same handful of vendors. What can't be bought is the organizational capability to put AI into the right process, the right decision rights, the right owner.

What does the new kind of leadership AI requires actually look like?

Someone with the authority and accountability to redesign how work gets done, not just sign off on software budgets. MIT Sloan calls this role the chief innovation and transformation officer.

Is Vietnam ready for AI transformation?

The workforce is ready. The organizations aren't, yet. Vietnam leads ASEAN with 39% of knowledge workers as proficient AI users, more than double the 16% global average, but leadership models are running behind that pace.

Key takeaways

• 91% of the AI transformation barrier is culture and leadership, only 9% is technology, so buying a stronger tool only addresses 9% of the problem.
• Technology is now something anyone can buy. What can't be bought is the organizational capability to put AI into the right process and the right decision rights.
• AI demands leaders who hold three real forms of authority: redesigning processes, changing incentives, and assigning accountability.
• Business leaders should flip their attention ratio from tools to people, put the right person in charge, and start with one process.

I'd genuinely like to hear how this plays out at your company. Is culture and leadership really the biggest barrier when you bring AI in, or is it something else? Am I missing something here?

Wishing you a steady hand through your AI transformation.

P.S.: if you're the only person at your company who believes in AI but you don't yet have the authority to change a process, what you need to fight for isn't a bigger tool budget. It's one of the three forms of authority above.