Written down, or it didn't happen
Architecture, decisions and trade-offs go in a document you own. If you replace me tomorrow, the next person picks it up without an archaeology project.
About
I spent nine years building the thing. Now I help you decide whether to.
Shivam Ghiya is a fractional CTO and AI strategy consultant based in Vadodara, India, working with founder-led companies in India, US, the UK and Australia. Nine years delivering ERPs, portals and integrations across healthcare, insurance, real estate, fintech and logistics.
Based in
Vadodara, India
Practice
OneTouchGrade
Working with
Clients in India, US, UK & Australia
I'm Shivam Ghiya. For most of the last decade I ran delivery: ERPs for insurance agencies, portals for real estate, intake systems in healthcare, apps for lending groups, and a lot of unglamorous integration work holding it all together.
The pattern got hard to ignore. The projects that failed almost never failed on engineering. They failed on a decision made months earlier: the wrong tool bought under pressure, a build that should have been a purchase, two systems integrated by whoever was free that week, an AI pilot with no baseline to measure against.
Nobody in the room was senior enough, or neutral enough, to stop it. The vendors certainly weren't, because they were paid to say yes. So I moved to the side of the table where I could.
That's what this practice is now. I work out how the business actually runs, design the system it deserves, then either build it or run the people who do, and stay long enough to keep it honest. Some engagements are three weeks. The good ones have lasted years.
What I won't do
- Take referral fees or vendor commissions, because the advice has to be worth something
- Recommend a build when buying is cheaper over five years
- Ship an AI feature with no baseline and no way to tell if it worked
- Leave a strategy document behind with nobody accountable for executing it
Ground covered
Six sectors. The same failure modes.
Domain knowledge matters, but less than people expect. What transfers is knowing which questions expose the real constraint.
Healthcare
Patient intake, records, care coordination
Insurance
Policy ERP, agent portals, claims workflow
Real Estate
Listings, lead routing, sales pipelines
Fintech
Investment platforms, encrypted partner APIs
Logistics
Bookings, driver allocation, automated invoicing
Accounting
GST e-invoicing, ledgers, role-based access
How I work
Four things I'll hold to, even when it costs me the work.
The cheapest answer wins
If a tool you already pay for solves it, that's the recommendation, even when the alternative is a project I'd be paid to run.
Small releases, measured
First release useful inside a quarter, with instrumentation from day one. Long silent builds are where budgets go to die.
I'll tell you when you don't need me
Plenty of calls end with "hire a good ops person instead." Saying that costs me an engagement and buys the only thing that actually compounds, which is a reputation for being right.
In their words
People I've worked for.
Shivam compiled an ERP portal for our insurance business, and as a result our own system provides everything, along with a customised dashboard.
Chirag VyasManaging Director, CV Financial Architect
Completed the work within the agreed timeline, with the must-have functionality, and stayed responsive to our queries through the whole development cycle.
Poonam Rohit GargFounder, Canvas2Craft
Great support from beginning to end. Genuinely informed, and went the extra mile at every stage. I would recommend them unreservedly.
Aaron AgarwalOperations Lead, Britna
Trustworthy, creative and professional. Exactly what you need when you have no technical background of your own.
Aparna MohanFounder, Soaked In Nature UK
Bring me the decision you keep postponing.
The AI tool you cannot evaluate. The CRM migration nobody wants to own. The build-versus-buy call that has been open for two quarters. That is the conversation.