The Macro: Web Forms Have a Completion Problem Nobody Talks About
I want to talk about something boring. Web forms. Specifically, the fact that they do not work nearly as well as everyone pretends they do.
The average web form has a completion rate somewhere between 20% and 40%. That means the majority of people who start filling out a form do not finish it. For longer forms, the kind used in insurance applications, medical intake, HR onboarding, and customer qualification, the completion rate drops further. Some multi-page forms see abandonment rates above 80%.
This is a staggering amount of lost value that businesses treat as an immutable fact of life. Marketing teams will spend thousands of dollars driving a potential customer to a landing page. That person clicks the “Get a Quote” button, sees a 15-field form, and closes the tab. The lead is gone. The ad spend is wasted. Nobody questions the form itself because forms are just how things work.
The form builder market is mature and competitive. Typeform made forms more visual. Jotform made them more flexible. Tally made them free. Google Forms made them ubiquitous. Fillout added database features. All of these products made incremental improvements to the same fundamental interaction model: present the user with a series of fields and hope they fill them all in.
The insight that forms could be replaced with conversations is not new. Chatbot companies have been pitching this for years. Drift, Intercom, and Qualified all built conversational lead capture products. But those products felt like chatbots, and chatbots have a trust problem. People know they are talking to a bot. The conversation feels scripted. The experience is often worse than the form it replaced because at least with a form you can see how many questions are left.
What has changed is that AI can now hold conversations that do not feel scripted. A large language model can adapt to context, ask follow-up questions based on previous answers, skip irrelevant fields, and handle unexpected responses without breaking. That is a fundamentally different interaction model than a decision-tree chatbot.
The Micro: Forms That Live Where Your Users Already Are
RowFlow replaces traditional web forms with AI-powered conversations that happen over text message, Slack, or embedded chat. Instead of sending someone a link to a form, you send them a text. The AI asks the questions conversationally, adapts based on responses, and captures the data in a structured format on the back end.
Alec Olesky and Jack Bubes are co-founders, based in New York, part of Y Combinator’s Summer 2025 batch. The founding team is lean and the product concept is focused, which is what you want to see from an early-stage company attacking an established category.
The channel strategy is the most interesting product decision. Putting the conversation on text and Slack instead of a website changes the dynamic in ways that matter for completion rates. A web form requires the user to navigate to a page, stay on that page, and complete the form in one sitting. A text conversation meets the user where they already are, allows them to respond at their own pace, and feels less like a chore.
Think about the specific use cases where this matters. Insurance applications. Medical patient intake. Employee onboarding surveys. Contractor qualification forms. These are all situations where the person filling out the form has low motivation and the organization needs high completion rates. The current solution is usually “make the form shorter,” which means sacrificing data quality for completion rate. RowFlow lets you keep the depth while changing the format.
The positioning as “the modern AI-native form” is deliberate and distinct from the chatbot framing. RowFlow is not selling a chatbot. It is selling a form that happens to be conversational. That distinction matters for buyer psychology. The people who buy form tools are not the same people who buy chatbot tools, even though the underlying technology overlaps.
The Slack integration is the piece I find most compelling for B2B use cases. Internal forms, the ones employees fill out for IT requests, expense approvals, feedback surveys, have even worse completion rates than customer-facing forms because employees have zero incentive to fill them out. Putting those interactions in Slack, where the employee already lives, removes the friction of navigating to a separate tool.
I could not find pricing information on the site, which suggests the product may still be in early access or sales-led motion. The green-heavy branding and clean Framer build communicate the right things for a productivity tool: simplicity and efficiency.
The Verdict
RowFlow is attacking a real problem from the right angle. The web form is one of the oldest interaction patterns on the internet and it has barely evolved in two decades. Conversational data capture is the obvious next step, and AI makes it viable in a way that scripted chatbots never could.
The risk is that form builders are deeply entrenched. Typeform alone has hundreds of thousands of paying customers. Switching costs are not high in dollar terms, but they are high in workflow terms. Every team that uses forms has built processes around their current tool. Replacing that tool requires retraining and re-implementation, and the ROI needs to be obvious enough to justify the disruption.
In thirty days, I want to see completion rate data compared to traditional forms for the same use case. If RowFlow achieves 70%+ completion where traditional forms get 30%, the product sells itself. If the improvement is marginal, the value proposition weakens.
In sixty days, the question is whether the Slack channel becomes a growth vector. If teams adopt RowFlow for internal forms first and then expand to customer-facing use cases, that is a land-and-expand motion that works. Internal use is low-stakes and high-frequency, which builds familiarity and trust.
In ninety days, I want to understand the data quality dimension. Higher completion rates are meaningless if the conversational format produces lower quality data. If the AI captures the same structured information that a form would, plus additional context from the natural conversation, RowFlow is not just replacing forms. It is making them better. That is the real unlock.