Abhishek Nayak keeps building the software behind the software

The business idea that stayed with Abhishek Nayak was not born in a brainstorm. It kept turning up as an expense. At Gharpay, the cash-collection company he started just after college, more than 150 people needed software to coordinate a physical operation. At Bicycle AI, his customer-support automation startup, the team needed interfaces for training data, inspecting models and running experiments. Different companies, different decades of technology, same bill: somebody had to build another internal tool.

These are the applications almost every growing company needs and almost no customer sees. The refund console. The sales lookup screen. The coupon generator. The panel that lets an operator correct a record without asking an engineer to touch the database. They are essential, repetitive and forever losing the priority contest to whatever the customer can see.

Nayak had been accumulating evidence about this invisible category for years. His longtime collaborator Arpit Mohan had been accumulating frustration. A backend engineer, Mohan disliked rebuilding interfaces in HTML and CSS. Their eventual third co-founder, front-end engineer Nikhil Nandagopal, disliked repeating the same patterns even though he knew how to make them. Appsmith came from the overlap: operational scar tissue, a technical annoyance and the recognition that the gap between a basic admin framework and a vast enterprise suite was wide enough to build in.

A founder before the job offer

Nayak studied biological sciences at BITS Pilani. The subject looks like a detour only if careers are supposed to arrive in straight lines. While at university he moved among entrepreneurship, public questions and climate work. He appeared on a youth panel at the 2007 Fortune Global Forum. In 2008, an essay he wrote about promoting the moringa tree through women's groups won third prize in an international youth contest. He later served as a youth delegate connected with the Indian Youth Climate Network at a UN climate conference.

By his third year, he had decided he wanted to start a company. He did not sit for campus recruitment. In January 2011, a month after graduating, he and Mohan founded Gharpay in Hyderabad. India's online-commerce boom still ran heavily on cash, and Gharpay offered merchants a specialized collection network. Nayak explained the decision in the language of temperament rather than destiny: “I knew that I would enjoy being an entrepreneur because I had very diverse interests and skills.”

Gharpay grew into a substantial field operation and was acquired by Delhivery. ClinkNow, another venture, was acquired by mobile-payments company Ezetap, where Nayak became director of product. Then came Bicycle AI, which tried to automate customer support and joined Y Combinator's Winter 2017 batch. It did not work out. Nayak has spoken plainly about shutting it down.

The sequence matters because Appsmith was not a first-time founder's elegant theory about how companies operate. It was the residue of operating. Cash logistics needed back-office software. AI support needed back-office software. At Accel, where Nayak worked as an entrepreneur in residence, his conversations with startups showed that nearly everybody had the same queue of neglected custom applications.

The launch that went nowhere

There is a version of the Appsmith story that begins with instant international adoption. It is true, but it starts too late. The team spent its first six months building and pitching the product. Nobody converted. The founders had to decide whether the audience was wrong or the product was not yet good enough.

They kept the audience. Developers, they concluded, simply held tools to a high standard. The team added features, improved quality and delayed the satisfying part of the story for roughly another year. When the next public launch arrived, a single announcement post brought teams from about 30 countries in the first week. The market had not suddenly learned to want internal tools. Appsmith had become usable enough to earn their attention.

“Think of us like WordPress, but for building internal user-facing applications.”Abhishek Nayak on Appsmith's original proposition

The analogy made a technical product legible, but Appsmith's choices were more specific than “low-code.” It was open source. It could be self-hosted. Developers could connect existing databases and APIs, arrange interface components visually, and still write JavaScript when a dropdown could not express the logic. The product promised speed without asking its core user to stop being a developer.

That bargain shaped distribution too. People could inspect the code, run it on their own infrastructure and show up in Discord when they needed help. Over time, users started answering one another's questions while the Appsmith team slept. Nayak's explanation of community is notably unsentimental: a group needs a shared interest or a shared love for the product. The useful behavior came first.

Money for the unglamorous layer

In June 2022, Appsmith announced a $41 million Series B led by Insight Partners, with participation from Accel, Canaan, OSS Capital and individual investors including Postman co-founder Abhinav Asthana and Cloudera co-founder Jeff Hammerbacher. The round brought disclosed funding to $51.5 million. At the time, the company said more than 7,000 active teams relied on it daily and its users had built more than 100,000 applications.

The capital gave the company time to deepen the open-source product and add the features larger organizations demand: security, governance, performance and support. The harder strategic task was to preserve the original clarity while moving upmarket. Internal tools cover a huge surface area. A small municipality, a fintech onboarding team and a support operation can all need custom software, but their workflows do not collapse neatly into one template.

Nayak has said roughly 60 percent of Appsmith use cases clustered around customer operations, including onboarding and support. The remaining 40 percent stretched from CI/CD utilities and sales demos to manufacturing, education and municipal work. The variety is the appeal and the trap. A platform must handle difference without becoming the same heavy system it was meant to replace.

Autonomy with a boundary

Nayak's approach to management follows a similar instinct: establish the frame, then leave room for judgment. He has written that Appsmith's founders did not set deadlines for the engineering team. Engineers and designers chose dates after agreeing on scope and the required quality. Managers tried to prevent scope from expanding during the build, and a missed self-imposed date carried no penalty.

He credited the practice with faster shipping and fewer anxious weeks around critical bugs. The interesting piece is not the absence of a date handed down from above. It is the exchange underneath: the makers get autonomy, while the organization protects the scope they estimated. Freedom is paired with a boundary.

Public recommendations describe Nayak as analytical, energetic and persistent. His own writing is more revealing in its plainness. He discusses the failed launch, the startup that closed and the recurring embarrassment of not having a personal website without sanding off the awkward parts. The voice is an operator's: here is the friction; here is the experiment; here are two questions for people who might use it.

The pattern moves again

Appsmith has expanded into AI agents for business workflows, aiming at the same operational terrain: support, sales, documentation and scattered company knowledge. More recently, Nayak has said he runs Appsmith and Kite, an agentic website manager. One idea he floated in 2026 was a button that could turn a LinkedIn profile into a maintained personal site - structure the story, select posts, provide a contact path and update when the profile changes.

It is early, and Nayak presented the idea as a test rather than a declaration. But it fits. A website goes live, then small changes become tickets, preview links, layout checks and nervous maintenance. The surface is new; the complaint is familiar. Somebody is repeatedly doing work that software ought to make lighter.

The line through Nayak's career is not a sector. Cash collection, payments, support AI, internal applications and websites do not share a tidy category. They share an operator's irritation with recurring coordination. His companies keep returning to the machinery behind visible work.

The quiet screens became the main event: the dashboard, the console, the form and now the website nobody has time to maintain.

There is a practical founder lesson in that path. An idea can survive a failed product, an acquisition and a change in technology because the underlying job remains. Nayak and Mohan encountered internal tooling at Gharpay. They met it again at Bicycle AI. Appsmith turned it into the product. Kite extends the question toward a public surface that still behaves like an internal burden for the teams maintaining it.

Nayak's stated aspiration for Appsmith has been straightforward: help engineers build useful applications for their teams without repetition and monotony. The words resist grandeur. They also describe a large amount of modern work. Every company is becoming a software company, the familiar phrase goes. Nayak's career adds the invoice in the fine print: every software company eventually becomes a builder of software for itself.