Most freelancers build their businesses around a collection of subscriptions. One tool handles scheduling, another creates estimates, and another takes care of invoices. Each solves a useful problem, but each also brings its own way of working.
I wanted to see what would happen if I built a tool around my own workflow instead.
The result is Pond: a private application for organising my freelance work, planning projects and creating invoices. I used Float as an initial reference, then shaped the product around the features I actually need.
The process was a useful exercise in product design and development. It also raised a bigger question: as custom software becomes faster and more accessible to build, how will businesses decide what to buy and what to make for themselves?
What Pond does
Pond brings the core parts of my freelance workflow into one place. I can:
Manage clients and projects, including budgets and day rates.
Plan work in a schedule, grouped by my availability or by project.
Build estimates using days or hours, then turn an estimate into a project.
Create invoices from project and client details and export them as PDFs.
Review planned work, project budgets and estimated value.
It’s designed for one person, so the interface reflects that. My availability and default day rate live in settings, and client colours carry through to their projects and scheduled work.
Just as useful were the things I chose to remove. Early versions included separate resource management, time tracking and utilisation reporting. As I refined the product, I realised those features weren’t serving the way I wanted to work. Because Pond is mine, I could simplify the app around that decision.
How I built it
I built Pond with Codex as my development partner, using React and HeroUI for the interface, PostgreSQL for the database, Google authentication and Vercel for hosting.
My role was to define the problem, make product and design decisions, test the app and direct each iteration. Codex helped implement those decisions across the interface, application logic and infrastructure.
The work moved between design and development throughout the sprint. I chose HeroUI’s Spotify theme with dark mode as the default, refined table surfaces and controls, added a custom app icon, and later brought the screens into Figma as editable designs with reusable components and variants.
I also set up Google-only sign-in so the deployed app is restricted to my account. Invoicing introduced another integration question: saving PDFs to Google Drive requires its own Google permissions and configuration.
This builds on the AI-assisted apps and web games I’ve been developing for a while. With Pond, I wanted to apply that process to something I’d use to run my business. In a short sprint, I took it from an initial brief to a deployed application on my own domain, then started bringing the interface into Figma for further refinement.
A different way to think about SaaS
Pond is a personal tool, not a replacement for a mature platform such as Float. Established products do much more than present a set of screens. They carry years of work in reliability, support, integrations, security and handling the edge cases that emerge across thousands of customers.
But the sprint changed how I think about the value of a subscription.
For a narrow workflow, it is becoming more realistic for an individual or small business to create something tailored to its own needs. A product that once required a significant development budget may now be approachable through a combination of good product thinking, design skills and AI-assisted implementation.
That creates a different kind of pressure for SaaS companies. The strongest reasons to pay may increasingly be the things that are difficult to reproduce and maintain: dependable service, specialist expertise, rich integrations, trusted support and functionality that has been refined through sustained use.
A tool built for one person can make different trade-offs from one built for thousands. That’s part of its appeal.
What this could mean for businesses
Most businesses already use third-party software for good reasons. Buying an established product can be the fastest and safest route to a reliable solution, particularly when a workflow involves sensitive information, complex integrations or many users.
At the same time, many businesses have smaller internal processes that sit awkwardly between existing products. They might be managed through spreadsheets, shared documents or an expensive subscription where only a fraction of the features are used.
Those workflows could become better candidates for custom software.
The question is no longer simply whether a business can build a tool. It also needs to consider who will maintain it, how access and data will be managed, what happens when an integration changes, and whether the workflow is important enough to justify that responsibility.
Pond gives me more control over my own process, but it also makes me responsible for the product’s upkeep. That trade-off matters.
Will more work move to one-person studios?
The same shift could affect how businesses commission digital work.
A project that once needed several people to move from an idea to a first working version may increasingly be achievable by one person with the right combination of design ability, technical understanding and AI-assisted tools.
That doesn’t make specialist expertise less valuable. It changes when and how it can be brought into a project. A one-person studio might lead the product, design and implementation, then involve specialists for areas such as security, complex engineering, accessibility or integrations.
The limiting factor becomes less about how quickly someone can produce an interface and more about whether they can make sound decisions, recognise weaknesses, test the result and take responsibility for maintaining it.
Pond is a small example, built for my own use. But it has made me reconsider both sides of the market: how much software a business needs to rent, and how much digital work a capable individual can deliver.
I don’t think every subscription will become a custom app, or every agency will become a one-person studio. I do think more businesses will have a choice they didn’t have before.
