At 13, Luis Loaiza taught a computer to hold a conversation. The software was primitive, the internet was noisy, and nobody had yet turned “AI agent” into a slide-deck incantation. What mattered was the sensation: words entered one side of a chat window and behavior appeared on the other. Two decades later, Loaiza is still pursuing that small piece of magic. The difference is that now the reply is expected to open a bank account, approve a credit application, collect a payment or sign a contract.
His career can look, at first glance, like a crowded browser with too many tabs open. There were Facebook games in Ecuador, an interactive-screen experiment, self-destructing messages, encrypted business communication, a last-mile logistics company built in Brazil, and finally Jelou AI, the conversational software company he leads from the United States. Yet the tabs all belong to the same session. Again and again, Loaiza has tried to reduce the distance between human intent and software action.
This makes him a useful character in the current AI boom. He did not arrive at conversation because a large language model made it fashionable. He arrived early, stayed late and accumulated all the inconvenient machinery that a talking interface needs if it is going to become more than a toy: identity, integrations, security, payments, logistics and the discipline of finishing what was started.
The useful obsession
Loaiza grew up in Guayaquil and studied computer science and multimedia systems at ESPOL from 2005 to 2009. By graduation, he had already helped organize the city's first Facebook Developer Garage. In 2010, his Club 40 multiplayer card game drew roughly 200,000 users in a month, by his own project account. It connected players across platforms at a time when the walls between devices felt much higher than they do now.
The projects became more serious without becoming less experimental. Qrayola turned phones into controllers for games on a shared screen. Hash made an opened message disappear after seven seconds. Criptext pursued secure communications for organizations. Each product asked a version of the same question: if communication is the interface, what else can the interface carry?
In 2012, Loaiza and Miguel Torres joined Start-Up Chile with Escapes With You, a travel marketplace. The marketplace changed shape. The pair had discovered that selling online in Latin America was only half a business; getting the purchase to the buyer was the more stubborn half. They went to Belo Horizonte and built Shippify around same-day delivery, verified drivers and software that coordinated the route.
Brazil supplied the sort of education that cannot be reduced to a certificate. Loaiza learned Portuguese, adapted to a market he did not know and later recalled the local cachaça and feijão tropeiro with equal warmth. Shippify grew from an experiment with a handful of drivers into a regional logistics operation. During the early pandemic surge, its monthly deliveries rose from 42,000 in May 2019 to 250,000 in May 2020.
“Technology alone doesn't build a company. Execution does.”Luis Loaiza, reflecting on Shippify
That line is the hinge in his story. Logistics punishes decorative technology. A parcel is either delivered or it is not. The route can be clever and the dashboard can glow beautifully, but somebody still has to arrive at the correct door. Loaiza carried that bias for completed work into conversational AI.
A chatbot walks into a bank
Jelou began in 2017 when a financial institution asked for help building a chatbot in Facebook Messenger. A basic assistant could answer a balance question. A useful one needed to reach the bank's systems, authenticate the customer, respect local rules and perform an action. The request opened a much larger problem than the original project suggested.
The company, co-founded with Alberto Vera, moved toward infrastructure that connects conversational channels to the systems where business actually happens. Generative AI made the conversation more flexible, but it did not remove the back-end work. In Latin America, every expansion brings another set of payment rails, identity sources, regulations and habits. The region's fragmentation is not a footnote to Jelou's product. It is the product brief.
Loaiza describes Jelou as his second logistics startup. It is an illuminating comparison. Shippify routed a package through drivers, streets and warehouses. Jelou routes an intention through models, databases, payment processors, identity providers and enterprise software. Both jobs are orchestration. Both are full of handoffs. Both fail when context disappears between one participant and the next.
The company calls its current territory transactional AI. The adjective matters more than the fashionable noun. A support bot can tell a customer how to apply for credit. A transactional agent can verify the customer, connect to a scoring system, present an offer, capture a compliant signature and record the result. Eloquence is pleasant. Completion pays the invoice.
The cold email and the long game
Loaiza is a software engineer, but his founder education includes a famous act of nerve. On April 16, 2018, anxious to find more capital for Shippify, he sent Jeff Bezos an email. Its opening was blunt: Bezos needed to know what Shippify was. The message eventually helped create a consequential relationship for the company. It is a good founder anecdote because it joins two forms of work that are too often separated: building something difficult and asking an important person to notice.
The same impatience with unnecessary distance appears in Jelou. Every redirect to a payment page, every repeated identity question and every handoff to a call center is a small confession that the software does not yet understand the whole job. Loaiza's published writing on AI Commerce argues for keeping orders, credit, payment and delivery updates inside one conversational flow. His phrase for the underlying habit is an “obsession” with removing friction between people and software.
In March 2023, after a leadership transition at Shippify, he took the CEO role at Jelou and put his full attention behind it. That year the company closed a $3 million seed round, rebuilt its platform and reported tripling its customer base. In January 2026, Jelou announced a $10 million Series A led by Wellington Access Ventures, with Krealo and Collide Capital participating. The capital is meant to expand the system behind those apparently simple chat exchanges.
“Users already live in messaging. We're making it the place where real operations happen.”Luis Loaiza on Jelou's computing thesis
There is also a less obvious continuity in the companies he has built: trust. A disappearing message asks whether communication can remain private. A delivery platform asks whether a stranger can be trusted with a package. A financial agent asks whether software can identify the right person, move money and leave a reliable record. Conversation may be the visible interface, but confidence is the hidden material.
Fit, not ornament
Loaiza's latest public framework is called AI FIT. He argues that companies should care less about how many AI tools they adopt and more about whether those tools make the organization faster, more efficient and more adaptable. It is an operator's argument dressed in biological language: fitness is proved by response, coordination and survival, not by ownership of impressive machinery.
That idea suits his own history. The Facebook ecosystem changed. Mobile platforms changed. E-commerce moved from novelty to infrastructure. Logistics scaled. Generative models arrived. Loaiza kept the core question and changed the implementation. His advantage is not clairvoyance. It is the compound experience of having watched conversations collide with real systems for years.
The most interesting future for Jelou is therefore not a more charming chatbot. It is a less visible transaction. The customer asks for what they need in ordinary language; the software carries the request across a complicated regional landscape; the job returns finished. No tour of the machinery. No scavenger hunt through five apps. The chat window becomes less like a receptionist and more like a control surface.
At 13, Loaiza was enchanted that a computer could answer. Now he is building toward the moment when an answer is merely the first, smallest sign that the computer understood. The real proof comes after: the account exists, the payment settled, the contract bears a signature, and the package is already on its way.