Luciano Kalil has a wonderfully inconvenient habit for a chief marketing officer: he likes to look at the work. Not the ceremonial version of the work, polished into a presentation and marched through a meeting. The live version. The campaign inside the advertising platform. The SEO curve in the analytics tool. The numbers before somebody has improved their posture. When his team meets, he would prefer an insight now to a deck three days from now.
This is less a management trick than a clue. Kalil is CMO of Pipefy, the process-automation company, but he arrived in marketing by way of code, retail systems, startups, online communities and product leadership. His career has changed labels often. Its central obsession has barely moved: how do you take complicated technology, arrange it around the way people actually behave and make useful work easier to do?
The question has carried him from the early Brazilian web to the current boom in AI agents. It also explains why his conversation about artificial intelligence sounds more like operations than sorcery. Agents may execute. Somebody still has to choose the goal, define the boundary, notice the exception and accept the consequence. Kalil’s future is automated, but it is not unattended.
The programmer who kept changing departments
Kalil’s published biography places him in business young, as a partner or co-founder from the age of 16. By 1996, his first company was already helping brands and people establish an internet presence. The web was small enough to feel like a neighborhood and strange enough that simply putting a company online could qualify as strategy. He worked with web agencies and projects for companies including IBM, Renault, UOL, Caloi and DuPont.
Then came a long education inside one company. At the Brazilian retailer Alô Bebê, Kalil entered as a programmer and left roughly a decade later as chief information officer. His own explanation of that climb was not poetic: work at the company as though it were yours. The reward was responsibility, followed eventually by the itch to own the thing for real. He called entrepreneurship “exciting and addictive,” which is a charming way of describing a professional preference for uncertain sleep.
His formal education mirrored the two halves of his work. He studied data processing, then completed a psychology degree at Universidade São Marcos. His focus included human-computer interaction, usability, collaborative environments, semantics and the gap between desire and need. Code told him what a system could do. Psychology asked whether anybody would want to do it.
A website builder before no-code got its wardrobe
In 2011, Kalil and Ricardo Cordts Monteiro began putting SitePX into the world. The software let small businesses build websites, blogs and online stores without learning a programming language. Today it would sit comfortably beneath a no-code banner. At the time, it was described more literally: a Brazilian website builder with hosting, search optimization, payments and design tools gathered in one place.
The founders reportedly put about R$1 million of their own money into the product and recovered the investment within a year. An early target was 9,500 completed pages. The platform later became an AWS Brazil case, attracted a reported R$3 million investment from Sul! Internet and, in Kalil’s publisher biography, was credited with serving 537,000 businesses.
The revealing moments were smaller. In a startup discussion group in 2012, users offered comments on a new SitePX landing page. Kalil thanked them, praised a customer’s inventive site and answered a technical question himself: no, the platform was not based on WordPress; it used PHP and jQuery. The founder was also the person in the thread.
He explained the product with a kitchen metaphor. SitePX could provide the cake recipe; the business owner still had to bring the ingredients. It was a modest claim for software, and therefore a useful one. Tools can remove technical friction. They cannot decide what a company knows, what it believes or what it should say. More than a decade later, the same division of labor sits beneath Kalil’s argument about AI.
From publishing pages to gathering people
Kalil’s next important product moved from the page to the group. He founded Duopana in 2017 as a platform for brands to create and manage digital communities. Social feeds rewarded the moment; communities, in his view, could reward continuity. A post on a broad network fell away as the feed refreshed. A focused group could accumulate knowledge, relationships and a reason to return.
That thesis became Community Hacking, the 2019 book he wrote with Marilvia de Oliveira. The book treated belonging as something that could be studied and designed across marketing, people management and technology. Kalil later contributed a chapter on communities to Trends: Mkt na Era Digital. He was not arguing that humans could be reduced to a funnel. He was trying to understand why some digital spaces invite participation while others merely collect accounts.
In 2020, Squid acquired Duopana. Kalil moved into the influencer-marketing company and eventually carried a striking bundle of responsibilities across product, customer experience and marketing. Those functions are often separated by organization charts and reunited by customer complaints. Having them under one executive made the loop explicit: what the product promises, how it feels and how the market hears about it are different views of the same experience.
He kept one foot in the startup classroom. Across years with Startup Weekend, Sebrae and Founder Institute programs, Kalil mentored founders on ideas, revenue models, bootstrapping, growth, experiments and traction. The work ranged across Curitiba, Londrina, Florianópolis and national cohorts. Mentoring is compatible with his public management style: experience is not a throne; it is useful material to place on the table, especially when one is willing to mention the mistakes that produced it.
The CMO who refuses to float above the work
Kalil joined Pipefy in 2024 and now sits on its executive team as CMO. The company’s product gives his career an almost suspiciously tidy destination. Pipefy organizes processes: requests, approvals, handoffs, documents, decisions and the minor bureaucratic weather systems that fill a working day. It now places AI agents inside those flows.
His marketing organization has been arranged to reduce some of the same friction. Product Marketing, Education, Community and Inside Sales operate with shared measures, bringing the story of the product closer to the mechanics of acquiring and teaching customers. Kalil remains willing to descend into an individual campaign. He has described opening tools during meetings and occasionally making a weekend round through SEO or advertising performance.
Hands-on leadership always carries a danger: attention can curdle into interference. Kalil’s stated balance is to look vertically at the delivery, offer pattern recognition earned through previous errors and preserve the autonomy of each specialty. The aim is not to seize every keyboard. It is to keep strategy in contact with evidence.
Where the meeting energy goes
This is why his apparently small objection to presentation work matters. A deck is a model of reality, and models become dangerous when their tidiness is mistaken for truth. Opening the actual tool lets the meeting encounter the stubborn detail. The conversion changed here. The handoff broke there. The customer did something no arrow anticipated. Reality, rude but efficient, has joined the call.
AI needs a skeleton
Kalil’s current writing on AI agents is built around workflow rather than spectacle. An agent should not be an isolated box performing a clever stunt. It should act at a defined point inside a visible process, with reusable components, access rules, monitoring and human checkpoints. Start with a small, valuable use case. Measure it. Correct it. Reuse what proves reliable.
The method is notable for how many limits it contains. Agents are not always the answer. Intermediate steps should be traceable. A person remains responsible for exceptions, approvals and critical decisions. Scale comes from modules that can be reused, not from creating another shiny agent for every inconvenience. The system improves because the organization can see it working.
This is SitePX’s cake recipe translated into enterprise operations. The platform can provide structure and automation. The company must bring intent, evidence, risk tolerance and context. In Kalil’s phrase, “AI agents execute. Humans give direction.” His aspiration is not a company without people. It is a company where people spend less time carrying repetitive work and more time defining what good work means.
Watch · CMO Agenda #18Luciano Kalil on practical leadership, workflows and AI agentsThere is continuity here, but not inevitability. A programmer does not automatically become a psychologist, author, founder, product chief and CMO. Kalil kept crossing the borders because each discipline exposed something the previous one could not explain. Technology could make a website. Marketing could bring someone to it. Psychology could illuminate the experience. Community could make them return. Operations could turn all of that into a repeatable result.
The useful future is less theatrical
AI has given business a fresh vocabulary for old temptations. Executives can buy a dramatic demo, declare transformation and leave the difficult middle untouched. But work lives in the middle: in permissions, incomplete data, judgment calls, handoffs and the person who notices that this case is different. Kalil’s career has largely taken place there.
That may be his most useful quality as Pipefy’s marketer. He can tell the technological story because he has built products. He can tell the human story because he studied behavior and spent years thinking about communities. He can test the operational story because he still wants the dashboard open. The message and the machinery are not allowed to become strangers.
So the scene that best captures Luciano Kalil is not a keynote stage or an executive portrait. It is a meeting with the live tool on screen. The numbers have not been groomed. A question lands, somebody clicks, and the team finds out. There are grander ways to imagine the future of work. Few are as likely to improve it before lunch.