Your AI budget has a people problem
Almost everyone is using AI. Almost no one is getting the return. The gap is not the tools. It is the skill to run them.
INSTAR
The paradox nobody wants to name
Nearly nine in ten teams now use AI in some form. About one in a hundred organizations would call their AI use mature. And most AI projects still show no clear return. Read those three numbers together and the story gets uncomfortable. The tools arrived. The results did not.
It is easy to blame the technology. Wrong model, wrong vendor, wrong prompt. But the pattern is too consistent for that. When the software is the same everywhere and the outcomes are all over the map, the variable is people. Whether the people using AI have the skill to turn it into something a business can actually bank.
That is the skills gap. It is quiet. It never shows up on an invoice. And it is the single biggest reason AI spend is not becoming AI value.
What the gap actually looks like
It is a manager who can write a decent prompt but cannot tell when the answer is confidently wrong. It is a team that automated a workflow nobody checked, then spent three weeks cleaning up the mess. It is a leader who bought licences for everyone and has no idea whether anything changed.
None of these are technology failures. They are judgment failures, data failures, and governance failures. The AI did exactly what it was told. The problem sat upstream, with the person holding the tool.
Two things have to be true at once
Close the gap and two things happen together. The organization turns AI into outcomes it can prove, and governs it without getting burned. The people doing the work move up the value chain instead of being pushed off it.
Those are not competing goals. They are the same investment. And they map onto two sets of skills a manager in the AI era has to hold at the same time.
The first set is the part AI cannot replace. Call it Management Excellence. Judgment under pressure. Leadership without authority. The ability to turn a pile of activity into a business result and defend the budget behind it. These were always the marks of a good manager. AI makes them rarer and more valuable, because once the machine handles the execution, what is left is the human call.
The second set is how you actually work with AI. Call it AI Mastery. It runs on four things. Data literacy, so you can judge whether the inputs are any good and read the outputs critically. AI fluency, so you use the tools well instead of poking at them. Orchestration, so you can direct AI and agents across a real workflow and measure what they add. And governance, so you handle risk, ethics, security and the EU AI Act before they handle you.
What AI can't replace
How to work with AI
FOUNDATION · PMI AI + DAMA-DMBOK STANDARDS
Seven capabilities, one manager
Put them together and you get seven capabilities. Ways of working. Power skills. Business acumen. Data literacy. AI fluency. AI orchestration. AI governance. The first three are the human core. The last four are the AI craft. A capable manager in 2026 is not strong in one column and blank in the other. They carry both.
Most people are lopsided. Strong operators who freeze at the word governance. Or AI enthusiasts who cannot connect a clever automation to a number the CFO cares about. The gap is almost always in the thin column. That is good news, because it means you do not have to rebuild a person. You have to find the thin column and fill it.
Everything here is anchored to real standards, the PMI AI Standard and DAMA-DMBOK, not to whatever was trending last quarter.
The seven capabilities in detail
Ways of Working
Management Excellence · foundation★Plan, scope, deliver. The project craft: decomposition, defining "done," managing risk. This is the discipline that makes directing AI possible in the first place.
The same habits now apply to a new kind of worker. Decomposition becomes breaking a goal into steps an agent can run. Defining done becomes verifying AI output. Risk management becomes setting guardrails. Without it, AI direction has nothing to stand on.
Power Skills
Management ExcellenceThe human core: judgment and critical thinking, creative problem-framing, leadership and developing people, communication and storytelling, empathy and trust, curiosity and adaptability, translating work into value.
These rise in value exactly as AI takes the execution beneath them. The World Economic Forum finds tasks needing empathy, creativity, leadership and curiosity are only about 13% likely to be automated. When fluency becomes common, judgment is what keeps a manager in the room.
Business Acumen
Management ExcellenceConnect the work to outcome, budget and ROI. Turn a clever AI use into a result a CFO will accept and fund.
This is the skill missing in the roughly 95% of AI projects that show no return a CFO can clearly link to the business. Usage is easy. Proving value is the constraint.
Data Literacy
AI Mastery · foundation★Judge data quality, governance, provenance, privacy and bias, and read model outputs critically. Management-layer data judgment, not data science.
The PMI AI Standard makes data quality a core principle: the impact of AI is only as strong as the data it receives. A manager who cannot judge the inputs cannot safely use, direct or govern the outputs. It is the substrate under everything else.
AI Fluency
AI MasteryKnow what AI is and is not. Prompting and context, human-in-the-loop, verifying output before you act on it.
This is the entry ticket, not the finish line. Half of all jobs paying over 100,000 dollars now require AI skills, up from a fifth in 2021. Fluency gets a manager in the door. It is necessary and no longer rare.
AI Orchestration
AI MasteryDefine scope and value, decompose work for agents, design AI-augmented workflows, supervise the agents, measure value and cycle time, own the outcome.
This is the thesis of the whole framework: orchestration over execution. As work shifts to agents, directing them well is where advantage is created, and at senior level the market has almost no supply.
AI Governance
AI MasteryMake AI safe and accountable: AI risk, decision rights, ethics, the EU AI Act, security and IP, human oversight and audit.
Responsible AI is among the fastest-growing skills in demand, and the EU AI Act's high-risk obligations become enforceable in August 2026. Governance moved from nice-to-have to legal requirement.
Two foundations hold the framework up: Ways of Working under Management Excellence, Data Literacy under AI Mastery. Neither wins on its own. Nothing above works without them.
Trajectory of skills
Judgment, critical thinking, decisions under uncertainty, leadership, influence, communication, empathy, creativity, value translation, risk and governance literacy.
Commercial and customer judgment, core financial literacy.
Coordination and status-chasing, routine drafting, routine reporting and data compilation.
Depth each role needs
Aware → Working → Advanced → Expert
| Role | Ways of Working | Power Skills | Business Acumen | AI Fluency | AI Orchestration | AI Governance |
|---|---|---|---|---|---|---|
| Project Manager | Advanced | Working | Working | Working | Working | Aware |
| Program Manager | Advanced | Advanced | Advanced | Working | Advanced | Working |
| Portfolio Manager | Working | Advanced | Expert | Working | Advanced | Advanced |
| PMO Lead | Expert | Advanced | Advanced | Working | Expert | Advanced |
| Functional / Line Manager | Working | Advanced | Advanced | Advanced | Working | Working |
| Executive / Director | Aware | Expert | Expert | Working | Working | Expert |
Read down a column to see who a skill is for. The heavy cells for senior roles, AI Orchestration and AI Governance, are exactly where the market has almost no supply.
Why these skills, and why now
For two hundred years a manager's value tracked the effort they could organize. AI ends that. When execution becomes cheap and nearly unlimited, effort stops being scarce and judgment becomes the constraint.
The evidence points one way. McKinsey finds 88% of organizations now use AI, yet only about 1% are mature and roughly two thirds are stuck in pilots. MIT-cited research finds around 95% of AI projects show no return a CFO can clearly link to the business. The World Economic Forum projects 170 million new jobs and 92 million displaced by 2030, and says that if the global workforce were 100 people, 59 would need retraining. Half of all jobs paying over 100,000 dollars now require AI skills, up from a fifth in 2021.
Read together they say the same thing. The money is being spent. The results are not showing up. The gap is not the technology, it is the management around it, and that is a skills problem. That is why these seven, and why now.
How INSTAR closes it
This is the whole reason INSTAR exists, and we work it from three directions.
For individuals, open courses take you along the path. Project foundations and PM Essentials for ways of working. Leading Without Authority and Judgment Under Pressure for the power skills AI cannot copy. AI Sprints and the Program Management in the Age of AI courses for fluency and orchestration. Start with the free skills check to see your own seven-capability profile and where your thin column is.
For certifications, there is a ladder that proves it. CAPM to start, PMP and PMI-ACP as you lead delivery, PfMP and PgMP at portfolio and program scale, and AI-specific credentials like PMI-CPMAI for running AI projects properly. The certificate is not the point. What it signals is.
For organizations, this is not a course you buy once. Our Transform advisory closes the gap as a system. We diagnose where your team actually sits across the seven capabilities. We build the program scoped to your context. We help you govern AI responsibly against real standards. And we embed the change so it holds after we leave. Corporate programs run both tracks at once, measured against your outcomes, not a generic syllabus.
Open courses
Walk the seven-capability path — from foundations to AI orchestration.
Certification ladder
CAPM, PMP, PMI-ACP, PfMP, PgMP and PMI AI credentials — level by level.
Transform advisory
Diagnosis, a program scoped to your context, AI governance, embedded change.
Where to start
If you take one thing from this, take the diagnosis. You cannot close a gap you have not measured. Most teams are one or two capabilities short of a real step change, and they are usually not the capabilities people expect.
Find the thin column. Then fill it.
SOURCES & STANDARDS▾
- PMI, The Standard for Artificial Intelligence in Portfolio, Program, and Project Management (2026) — eight principles, five performance domains, the AI life cycle, and data quality as a core principle.
- DAMA International, Data Management Body of Knowledge (DMBOK2) — the DAMA Wheel of eleven knowledge areas with Data Governance at the core; CDMP credential (~13,000 holders); DMBOK 3.0 in development for AI.
- McKinsey — The seven operating truths of AI-native companies (2026); How European organizations can treat skills as a strategic priority (2026); the agentic organization; European competitiveness research (skills gap worth 500 billion to 1 trillion euros a year by 2030).
- World Economic Forum — Future of Jobs 2025; New Economy Skills: Unlocking the Human Advantage (Dec 2025): empathy, creativity, leadership and curiosity ~13% likely to be automated; 170M new / 92M displaced jobs by 2030.
- Coursera — Industry Skills Brief 2026: critical thinking a core competency (triple-digit growth); Responsible AI and information privacy among the fastest-growing skills.
- Accenture — The Age of Co-Intelligence (2026).
- MIT-cited research — ~95% of AI projects show no return a CFO can clearly link to the business.
- Ladders — half of jobs paying over 100,000 dollars now require AI skills, up from a fifth in 2021.
- EU AI Act — high-risk obligations enforceable from August 2026.