O9X didn’t start as a product idea. It started as a control problem.
In a manpower services business, payroll and invoicing solve the visible issues. Once those are in place, salaries go out on time, invoices get raised, and the business looks stable from the outside.
We had that covered.
We were using Zoho Books for accounting and a proprietary payroll system we’d paid ₹3 lakhs to license. On paper, this should have eased a lot of issues within my business.
It didn’t.
What it eliminated were calculation mistakes. What it didn’t eliminate was operational ambiguity.
When Systems Go Silent
As we scaled, customer complaints became routine. Small issues. Repeated issues. Issues my ops team swore hadn’t happened.
Every complaint triggered the same manual ritual:
- checking WhatsApp threads
- opening Excel trackers
- calling field supervisors
- reconstructing events from memory
There was no single source of truth. Just people, interpretations, and partial records. There was no immediate financial loss. But the mental load was real. And it was obvious this setup could survive at today’s scale—yet would quietly collapse as complexity increased.
Payroll systems don’t fail loudly. Operations do.
The Experiments That Failed
Before deciding to build, I tried to fix this the “right” way: exploring other tools I could buy off the shelf. I didn’t want to waste days on AI coding (which rarely works for complex projects anyway). I wanted to use what already existed.
I ran two experiments. Both failed. Both were instructive.
1. The No-Code Trap (Airtable)
I started with Airtable workflows. We automated small wins, like emailing customers when a field executive visited a site.
What worked
Customers liked the transparency. Deployment was fast.
What broke
The system became brittle. Costs scaled faster than value. One small change required extensive reworking and sometimes even external paid tools.
Verdict
Useful for prototypes and small internal workflows. Too costly to scale for core operations.
2. The “Everything” ERP (ERPNext)
Next, I evaluated ERPNext. It’s powerful. Comprehensive. Open-source
What worked:
Coverage and robustness.
What broke:
The hidden cost. I had to restructure my business to fit the software.
I hired external developers. I spent time and money. The failure wasn’t technical—it was transactional. No external team understood the nuance of Indian manpower operations. Explaining the business became harder than building the system myself.
Verdict:
Capable tool. Prohibitively time-consuming and expensive to customize.
The Legacy Prison
While these experiments failed, our existing stack was bleeding us quietly. Zoho Books is doing a great job as it gave us cloud visibility and reduced dependence on individuals.
The payroll software did not.
We paid ₹3 lakhs to buy it. We still pay ₹35,000 every year for mandatory support.
It runs offline.
It lives on a single desktop.
Moving it to the cloud would cost another ₹50,000 yearly.
In a business where operations happen in the field, my largest cost center is locked inside one machine. If that desktop goes down, I’m blind.
The Decision: Don’t Replace. Bridge.
At this point, the choices were clear:
- Airtable was too fragile and costly at scale
- ERPNext was too rigid
- Legacy payroll software was too archaic with high recurring costs
So I chose the third path.
I’m building O9X—an internal Ops ERP, built on mature open-source tools, with a hard infrastructure ceiling of ₹4,000 per month.
The strategy is not to replace everything. The strategy is Frankenstein.
The Integration Plan
Phase 1: Field Truth (Input)
I need to end “he said, she said.” Mobile-first capture of attendance, shifts, and site visits. No Excel trackers. No WhatsApp screenshots as evidence.
Phase 2: Payroll Kill Switch (Engine)
Once operational data stabilizes, payroll logic moves into O9X. We’ll run it parallel to the legacy system for a month. If the numbers match, the decision is final: the legacy system goes, the offline risk goes, and the ₹35,000 support tax goes with it.
Phase 3: The Zoho Bridge (Output)
O9X will never touch the balance sheet. Zoho will. Clean invoices and payroll entries flow into Zoho via APIs. No need to build a complex system with too many moving parts. Let the experts at Zoho handle that.
This Is a Build Log
O9X is not a generic ERP. It is not a SaaS experiment. It is not something I’m building to sell.
It is a survival mechanism for a low-margin, high-volume business.
I’m building it to own my failure modes:
- I control how data enters the system
- I decide which errors are blocked
- I encode business quirks instead of explaining them repeatedly
This series will document the build—the stack, the schema, the failures, and the rewrites.
I stopped buying software because it kept hiding risk. I started building because I needed to see it clearly.