See your company's AI maturity level and your teams'. 28 questions, 20 minutes, no signup.
89% of companies use AI. 6% capture meaningful impact in their results. The distance between those two numbers is not technology. It is where the company moved and where it did not. In 20 minutes you find out where yours stands, against whom, and what to do in the next 90 days.
{{ p.t }}
The two studies say the same thing along independent paths: training people without touching the process does not move the result. That is why this diagnostic looks at four places at once, and not only at how well your team can use the tool.
Esquadro is the AgenticPeople method for installing AI culture inside a business area. It starts from two questions anyone understands: is it inside or outside? is it one person or everyone? That crossing gives you four places where change happens. Either all four, or it does not hold.
{{ q.d }}
{{ q.d }}
{{ blocoD }}
{{ rLead }}
{{ firmeTitulo }}
{{ firmeEvid }}
{{ k.d }}
Each quadrant fills from the bottom up depending on how much of that place is already installed. The marked quadrant is the primary bottleneck: the emptiest one, and the one the plan starts with. The dark side is what nobody sees. The bright side is what gets measured.
He and Puranam describe four levels and are explicit: level 4 cannot be decreed. It emerges from mastering the levels below. Nobody skips a step.
{{ p.d }}
{{ gargaloGap }}
{{ gargaloMove }}
{{ benchLead }}
{{ pLead }}
{{ f.cl }}
Nothing moves up a level without proof. Individual proficiency only becomes a team thing when it becomes an artifact someone else runs with the author nowhere near. A team artifact only becomes a process when it has a named owner, a metric on the dashboard and passes the data standard. Without those two ratchets, the program is training with a nice name.
Four quarterly blocks. Each one closes something before the next opens. The fourth ends with the change repeating with nobody from outside in the room.
{{ f.cl }}
Your company can run this plan without AgenticPeople. It was designed that way on purpose. But it would be dishonest not to say where it usually gets stuck.
{{ h.b }}
The diagnostic asks for no signup. The report does, because a copy also goes to your email.
Use the Save as PDF button in the tab that opened. A copy also goes to the email you left.
The portrait is yours and so is the reading. If you want to read it together, the 30 minutes stay open, no strings.
{{ n.v }}
See your company's AI maturity level and your teams'. 28 questions, 20 minutes, no signup.
89% of companies use AI. 6% capture meaningful impact in their results. The distance between those two numbers is not technology. It is where the company moved and where it did not. In 20 minutes you find out where yours stands, against whom, and what to do in the next 90 days.
Organizational factors: culture, manager support, people practices. They explain more than twice the reported AI impact compared with individual factors like mindset and behavior.
Among those who capture results, 73% fundamentally redesigned their workflows. Among the rest, 25%. They are also 3.3× more likely to pursue business transformation.
The two studies say the same thing along independent paths: training people without touching the process does not move the result. That is why this diagnostic looks at four places at once, and not only at how well your team can use the tool.
Esquadro is the AgenticPeople method for installing AI culture inside a business area. It starts from two questions anyone understands: is it inside or outside? is it one person or everyone? That crossing gives you four places where change happens. Either all four, or it does not hold.
Judgment to decide what to hand to the machine, what to check and what never leaves your hands. Critical sense in front of convincing and wrong output.
Observable proficiency: a project instruction, attached context, an automation built, an agentic tool run end to end.
What can be said at no cost, what happens when the agent gets it wrong, and the agreement about what gets automated and what never does.
Use policy, data protection, a named owner per flow, an audit trail, proprietary context and the redesigned process.
With the primary bottleneck named: the weakest place, which is where the plan starts.
Base Cetic.br · TIC Empresas 2025, 4,174 companies, fieldwork Feb 2025 to Jan 2026.
He and Puranam, INSEAD 2026. And why you cannot skip a step.
With an owner, a deliverable and a closing criterion per phase. Plus the 12 month arc in four blocks.
Microsoft · Work Trend Index 2026 (n=20,000 across 10 markets, Brazil included, fieldwork Feb 18 to Apr 7, 2026). 67% × 32%, 26% say leadership is clearly aligned, managers who use AI in the open raise perceived value by 17 points. McKinsey · The State of AI 2026, “On the road to ROI”, published Aug 25, 2026 (n=1,719 across 97 countries, fieldwork May 4 to Jun 8, 2026). 89% use it, 37% report EBIT impact, 6% are high performers, 73% × 25% on workflow redesign. Cetic.br | NIC.br · TIC Empresas 2025 (n=4,174 Brazilian companies, fieldwork Feb 2025 to Jan 2026). 17% of Brazilian companies use AI, 15% among small ones, 50% among large ones. Vivianna Fang He (UCL) and Phanish Puranam (INSEAD) · The AI Maturity Pyramid, INSEAD Knowledge 2026.
The diagnostic uses Esquadro, a proprietary AgenticPeople method. The architecture of the four places comes from Integral Theory. The sequence of the phases comes from Theory U. The two ratchets come from Kegan's subject to object transitions. None of those origins needs to show for the method to work.
This diagnostic runs entirely in your browser. No answer is sent, stored or passed to anyone, including AgenticPeople. When you close the page, it all disappears. If you want to keep it, use “Save as PDF”.
This is an orientation instrument, not an audit. It does not replace legal advice on data protection, an information security assessment or technical due diligence. The recommendations come from your own answers. If they are optimistic, the portrait will be optimistic.