Gondola Doubles Revenue With Codex as Non-Engineers Ship Code

Creative-careers platform Gondola more than doubled its revenue year over year without adding payroll costs, while its lead engineer's monthly shipping output rose 50% through the OpenAI coding agent Codex. Across the whole engineering team, output climbed 25% with no change in headcount. The figures, reported by the company, come from a small startup that has handed product prototyping power to non-engineers — and watched both velocity and revenue grow.

The stakes extend beyond one company. Gondola's experience is a concrete data point in the broader debate over whether AI coding agents compress the gap between what small teams imagine and what they can actually ship, shifting product work away from engineering bottlenecks and toward the people closest to customer problems.

Where the numbers come from

Gondola reports a cluster of results since adopting Codex in early 2026:

• 25% increase in monthly engineering output across the team, measured with no change in engineering headcount.
• 50% increase in the lead engineer's merged changes per month.
• 2x company revenue year over year, with no added payroll costs or significant overhead.
• 2.5x growth in creators hired by brands through the platform year over year.

The team is also shipping larger changes, according to the company, while pull-request review time has held steady. Those are self-reported figures from a case study published by OpenAI, and Gondola's scale — a small team — means the percentages rest on a narrow base. But the pattern is consistent: more output, larger changes, flat review time.

What is Gondola?

Jared Kleinstein founded Gondola after running the creative production agency Fresh Tape Media, where he noticed that exceptional creative work was being made for social media while the makers received little credit. Gondola's platform lets creatives show what they built, prove its impact, and convert that record into their next job or brand deal.

For a small team, the binding constraint was engineering capacity. Small experiments competed with core product priorities for limited engineering time, and even routine information-finding and drafting required sifting through multiple tools and systems.

How Codex changed the workflow?

Gondola's product ideas come directly from customer conversations and operational problems that the business development and leadership team sees firsthand. Before Codex, that team would define a need, write requirements, and then wait for an engineer with free capacity to build a first version.

Now the same team uses Codex Cloud — the cloud-hosted version of the coding agent — to explore the codebase, prototype an interface or workflow, test assumptions, and often reach a functional first version before an engineer steps in. Kleinstein and Jojo Hesketh, Gondola's Creator and Client Partnerships Manager, have both merged their first-ever code contributions into the main codebase.

"A product manager doesn't need to stop at a written product requirement anymore," Kleinstein said. "They can bring engineering a working prototype."

Kleinstein framed the shift in ambition, not just speed: "Using OpenAI has increased the ambition of what we think a company our size can build."

Gondola's engineers, for their part, use Codex through the command-line interface (CLI), editor tools, and GitHub pull-request review. The company connects Codex to its repositories, development environments, and Slack, so product context can inform development. Engineers remain responsible for architecture, security, data handling, and production readiness — a division of labor the company has kept explicit.

What did the team actually build?

The clearest example is Programs, a tool Gondola built to help companies find creatives, review applicants, manage projects, and track budgets, with plans to add contract coordination and payments. Hesketh sponsored Programs and used Codex to prototype its workflows before handing them to engineers to refine and release.

The team started the work with GPT-5.4, did most of it with GPT-5.5, and used GPT-6.1 for later features, optimizations, and bug fixes — a model progression that tracks how quickly the underlying tools iterated during the project. Kleinstein estimates Codex saved days, if not weeks, of development time and helped the team launch Programs with fewer bugs than hand-coding would have produced.

"The Programs feature benefited from our ability to rapidly prototype, throw away things that didn't work, and keep the things that did," Kleinstein said.

The second example was time-sensitive. Before Codex, opportunistic ideas might sit in a backlog awaiting engineering resources while the window closed. This summer, Gondola built and tested a World Cup Creator Map while it was still timely: a map letting brands, teams, and agencies find photographers and videographers in World Cup host cities, built rapidly from data Gondola already held. Hesketh sponsored the project and built a fully working prototype with Codex; engineers polished it for launch shortly after. Kleinstein describes the build as "incredibly quick," crediting its small, self-contained scope and hard deadline, and believes both the Creator Map and Programs shipped with fewer bugs and more polish than they would have without Codex.

Where does ChatGPT fit in?

Building product is only part of running the company. Gondola also uses ChatGPT to prepare proposals, manage schedules, find information across company systems, and automate operational tasks.

The most developed example is a daily report template employees use each morning to plan their day. It integrates Google Workspace — Gmail, Calendar, Sheets, and Drive — plus Asana, Slack, QuickBooks, weather information, and other sources into a single view of the day ahead. Kleinstein says preparing such a report manually would take "hours and hours of work," but it assembles quickly in the ChatGPT setup.

One guardrail remains: a person still reviews consequential work before it goes out externally or feeds a business decision.

Why does this matter beyond Gondola?

The numbers Gondola reports — doubled revenue, 2.5x creator hires, 25 to 50 percent output gains on a flat headcount — sketch a plausible template for small software companies in the agent era: business staff prototype, engineers own architecture and production readiness, and the bottleneck moves from implementation capacity to judgment about what to build.

"OpenAI gives product people more ability to build, engineers more time for difficult engineering, and our business team more time to spend with customers, partners, and creatives," Kleinstein said.

Looking ahead, Gondola wants AI to take over more of the coordination work that connects companies and creatives — reducing the overhead of finding talent, managing budgets and approvals, handling contracts, and getting people paid through Programs. The company's stated aim is to keep expanding what a small team can deliver in service of its founding mission: helping creatives get credit for their work and turn it into their next opportunity.