Pathao Tickets: replacing Zendesk with a patched open source helpdesk

Pathao's customer support ran on Zendesk. It cost about $1,000 a month, and it only covered tickets.

Support wanted two more things: live chat, so customers could get an answer without waiting on an email thread, and automations, so the same kinds of tickets did not need a human to sort them every time. We asked Zendesk for a quote. The number for chat and automations at our volume was huge, a multiple of what we were already paying.

The question that started it

Our CTO asked a simple question: can we have something simple instead?

Not a rebuild of Zendesk. Something that did the three things support actually used every day, and cost a fraction.

The suggestion

I suggested we stop looking at vendors and pick a solid open source helpdesk, self-host it, and patch it until it fit how Pathao support worked.

The reasoning was plain. Ticketing, an agent inbox and a chat widget are a solved problem. Good open source projects already do them well. What no project does out of the box is your company's routing rules, your internal tools and your edge cases. That last part is a much smaller job than building a helpdesk, and it is the only part that is really ours.

What we built on top

We called it Pathao Tickets. On top of the open source base we added or reworked:

• Ticketing that matched how support teams were split, so a ticket landed with the right team instead of a shared pile.
• Live chat for customers, feeding into the same agent inbox as tickets, so agents worked one queue, not two tools.
• Automations based on tags. A ticket gets tagged, by a rule or by an agent, and the tag drives what happens next: which team owns it, what priority it gets, which canned flow kicks in. Support leads could reason about tags. They could not reason about code.

Most of the work was monkey patching: small, targeted changes to the open source code instead of a long-lived fork that drifts. Each patch had one job. That kept upgrades possible and kept the surface area small enough for a few engineers to own.

The result

Pathao Tickets handled a very large volume of support conversations, and it ran for around $300 a month in infrastructure.

That is less than a third of what Zendesk cost for tickets alone, and it included the chat and automations Zendesk wanted a much bigger number for.

What I took from it

• Price the vendor quote against the boring version. The question was never "can we build Zendesk". It was "what do our agents actually click every day". That list was short.
• Buy the solved problem, own the specific one. Open source covered the inbox and the widget. Our time went only into routing, tags and integrations, the parts that were specific to Pathao.
• Patch small, not fork big. Small, single-purpose patches are cheap to review and cheap to carry through upgrades. A big fork turns into a second product you never meant to maintain.
• Give operators a lever they understand. Tags made automation something support leads could change themselves. That is why the automations actually got used.

If a SaaS quote is pushing you toward a number that makes your CTO wince, check what your team really uses. Often a well-maintained open source project and a handful of careful patches cover it for a fraction of the price.