From a broken job board to 7,000 paid subscribers in two months

The problem: a job board that looked alive but wasn't

The client works with a large professional community in the education industry. They already ran a job board for that community, but it was broken. Jobs went out of date, apply links led nowhere, and nothing was automated, so every new listing meant someone adding it by hand.

A job board lives or dies on trust. If a member clicks three dead links, they never come back, and they certainly don't pay for a subscription. The client didn't need a patch. They needed the board rebuilt as a product.

Planning the whole platform in Ideaunch

Before anything was built, Deepesh took the project into Ideaunch. It mapped the whole platform as a flowchart: what members see, what runs in the background, and how the plans work, from a free tier that shows part of the board to a full membership that unlocks everything.

Ideaunch then laid out the prototype and the UI/UX for every page: the member dashboard, the full job board with search and filters, a matched-jobs page, a companies directory, an application tracker, onboarding, job alerts and the AI tools. It also set the build timeline, so the client knew what would ship and when.

The exact prompts, the right tools

For each page, Ideaunch wrote the exact prompt to paste into the builder. It also picked the stack: Lovable Cloud for the app and the database, Firecrawl to read job pages from employer and career sites, Zapier to connect the pieces, and GHL (GoHighLevel) for checkout, subscriptions and member emails.

Because every prompt came from a finished plan, the build moved in a straight line instead of looping between broken versions.

What got built: an automated AI job board

Today the board runs itself. Automated scrapers read the list of employer and career sites, sort each job into the board's categories, check that the apply link really works, apply the location rules and publish. A watcher agent keeps quality high by removing jobs that don't fit the board, so members only see live, relevant roles.

Matching is the heart of the product. Every job and every member profile gets an AI embedding, a kind of meaning fingerprint, refreshed every hour in the background. The skills members already have are translated into the business skills employers search for, so they find roles they would never have searched for themselves. Each match comes with one to three plain-English reasons, and a "Not for me" button that hides the job and retrains that member's ranking instantly.

Speed was rebuilt too. The old board downloaded about 11,000 jobs into the browser and took 18 to 37 seconds to load. Now search, scoring, counts and paging happen inside the database, at about 17 milliseconds per page.

Payments, emails and operations on autopilot

GHL handles checkout and plan changes through webhooks, with a provisioning queue so a paying member is never stuck on the free plan. Transactional emails, weekly job digests and new-match alerts go out on their own. An admin suite shows traffic, trial and membership journeys, conversions and search trends, and a watchdog alerts the team if analytics, email or tracking ever break.

The result: 7,000 paid subscribers in two months

The new board launched two months ago. Deepesh delivered it as a service, and the client now sells access to their members as a subscription. It already has 7,000 paid subscribers, with jobs, payments and emails fully automated.

Ideaunch's job didn't end at launch. It now tracks the board's analytics and shows the client which channel to run ads on, so growth money goes where members actually come from.