The first customer paid $300 a month. The work was search optimization for a porta-potty rental company. It is the sort of detail that makes a founder story feel as if somebody left a joke in the minutes. For John Phillip Morgan, better known as JP, and his partners Dave Rogenmoser and Chris Hull, it was ordinary early work: a small client, an actual invoice, and three people learning whether they could make a business together.
Jasper, the AI marketing company they founded years later, arrived to a very different reception. By October 2022, it had announced a $125 million Series A. Yet the founders' own account puts the beginning long before the financing, before the AI product, and before a market eager to hear the words “generative AI.” They had already tried software companies, run a marketing agency, taught marketing courses, and built a product called Proof. Morgan kept showing up at the technical end of each new idea.
That history matters because the engineering lesson at the center of Morgan's Jasper story was learned when a previous product became hard to change. Proof had grown, entered Y Combinator, and raised money. It had also accumulated technical debt that boxed its makers in. Jasper's architecture was a deliberate answer. An AI product would need to move while the technology, customers, and even its own purpose were moving.
A partnership with a peculiar first customer
The trio describe an early pact in their Y Combinator account. They wanted to start a company, earn enough each month to support their families, and place their friendship and commitment to one another above either goal. That last clause sounds sentimental until the list of ventures starts to grow. Partnerships are easy to admire in a launch photograph. They are harder to sustain through the versions of a company that never quite become companies.
They shut down two SaaS attempts, took on marketing work, and landed the porta-potty account. The agency then gave way to a course business teaching marketing. For about three years, they learned the subject by selling and teaching it. That gave them a close view of the everyday problem that later informed Jasper: marketers needed more copy, more variations, and more ways to test what worked. An engineer in that setting learns what a customer asks for after the demo ends.
Morgan's public career record also lists Payfunnels, where he was a co-founder and CTO before Proof. These names matter less as a string of ventures than as evidence of a habit. The partners would abandon one shape of business and keep the working relationship. It made the arrival of a new technology less like a cold start and more like a new problem for a team that already knew how to argue, build, and return to the table.
Proof's success, and its expensive constraint
Proof displayed recent customer activity on websites, turning the idea of social proof into a software product. The team applied to Y Combinator and was rejected. The feedback, according to its founders' account, was that the idea looked too much like a feature. They kept building, grew the business, applied again, and were accepted into the Winter 2018 batch. A week after Demo Day, they had raised $2.2 million at a $12 million cap.
The rejection is a good detail because it complicates the tidy version of the story. The team did not receive a clear verdict from the market and obey it. They gathered more evidence, changed the pitch, and won a place in the accelerator. They also moved to Austin. Proof was real enough to hire people and earn revenue. The product, however, became increasingly difficult to extend.
Rogenmoser later described a development trap: the application had gathered so much technical debt that new work seemed to demand a larger rewrite. When growth stalled, the company could not easily build its way toward another future. Morgan, its CTO, carried that experience into the next venture. Rogenmoser credited him specifically for changing the engineering approach at Jasper. Modules were made easier to replace. If one part of the application became obsolete, the team could rebuild that part quickly.
“We built Jasper to be more componentized, so product modules were easier to switch out.”Dave Rogenmoser on the lesson Morgan took from Proof
It is a modest sounding design choice with large consequences. Many startup stories treat speed as a personality trait: the founders were bold, restless, or unusually decisive. Morgan's work makes speed tangible. It lives in the arrangement of code and the cost of changing one piece. In a field where the underlying language models were improving rapidly, the ability to replace a component was a business option.
The text that became a product
After Proof stalled, Rogenmoser returned to marketing consultancy and coaching. One lesson in his course covered ad copy. Students had trouble writing it well, and he wondered whether GPT-3 could help. Through Y Combinator contacts, he had access to the model. He texted Morgan about building a simple interface on top of its API. Morgan made a working MVP in about a month.
The scene has none of the theater later attached to AI. A copywriting teacher had a problem in front of him. An engineer could make a usable tool quickly. They showed it to people in the coaching program, and the reaction gave them reason to continue. Proof had taught them the cost of building too far ahead of demand. This time, a small product met users while it was still small.
The earliest idea centered on creating and testing ad variations. It expanded to long-form writing and other content as the team saw more possible uses. Founded in Austin in January 2021, the company took several names on its way to Jasper. Its first public product translated the power of a language model into fields and tasks that a working marketer could recognize. A person who needed an ad headline did not have to begin with a blank prompt and a research project about AI.
Morgan's contribution is easy to miss precisely because it is the part meant to disappear. The interface put a complex model behind an ordinary workflow. The modular build meant the team could keep changing the workflow. The achievement was less a single algorithm than a product that made emerging technology usable before the broad public had a settled vocabulary for it.
A small team meets a large market
Jasper's first year produced numbers that seem oddly mismatched. Rogenmoser said the company ended it with about $35 million in annual recurring revenue, nine employees plus two support workers, and roughly 40,000 customers. He described heavy use of Zapier automation and a decision to hire slowly. Morgan's technical design sat inside that unusual operating model: a small group maintaining a product whose users arrived much faster than an ordinary startup team could expand.
By October 2022, Jasper announced its $125 million Series A at a reported $1.5 billion valuation. Built In Austin reported more than 70,000 paying subscribers at that time. The photo from the announcement catches Morgan with Hull and Rogenmoser against a blue backdrop. Morgan is holding a keyboard, the closest thing the portrait has to a prop from his actual work. A small robot appears too; magazine photographs have their own understanding of software.
The growth did not freeze the product's identity. Jasper added an extension that brought its writing assistance into other places people worked. Later it moved toward tools for marketing teams and businesses that needed consistent brand voice and shared context. Its current description calls it a marketing agents platform, built for campaigns, personalization, and other workflows. The trajectory is visible in the product itself: an early copy helper grew into a broader system for organized marketing work.
There is a useful restraint in looking at those figures. A funding round measures investors' willingness to back a company at a point in time. A customer count measures a different thing. Neither tells us Morgan's personal wealth, and neither explains exactly how he spent every day. They do show the demands on a technical founder whose architecture had to bear a fast-changing product and a suddenly large audience.
Make it possible to replace the piece that no longer fits. The next turn may arrive before the current one is finished.
What remains visible now
Morgan keeps a lower public profile than the funding announcements might suggest. His GitHub description is plain: entrepreneur, software engineer, and Jasper co-founder. His personal site calls itself “Notes on software” and says it is about products and building durable software. It lists Payfunnels, Proof, Y Combinator, Jasper, and a 2025 project called Safelead. In 2026, it also links to notes and experiments on web apps, APIs, and software agents.
One former colleague's LinkedIn recommendation praised his commitment, creative solutions, and attention to relationships with coworkers. It is a narrower and more credible glimpse of personality than any sweeping founder label. The founders' own account supplies another: long before Jasper, they chose to keep working together and to make that relationship an explicit goal. Whatever else changed, the trio remained a trio through very different businesses.
Morgan's later writing brings the story into a neat, though unforced, circle. The engineer who helped build Jasper so it could survive the next change is writing about durable software while software itself changes again. The tools on his site are newer than Jasper's first prototype. The practical question feels familiar: what can a builder make today that will still be useful when the surrounding technology moves?
The $300 client and the nine-person AI company belong in the same biography. One explains how Morgan learned the market with his partners; the other shows what happened when their next product reached it. Between them sits the less photogenic fact that makes the story his: he had seen a successful application grow too rigid, and when the team got another chance, he helped build room for the next turn.