// Custom apps and integrations
Enter it once. Every system agrees.
I connect your CRM, accounting, store, warehouse and scheduling so a change in one shows up in the rest. A failed sync lands in a queue someone sees, with the reason next to it.
First step: a free 15-minute call. Then a sprint from $7,500: a written plan and a fixed price for the build.
System integration connects the tools a business already runs so each record is entered once. Eyal Gantz sets one owner for every field, syncs both ways where that is safe, and sends every failed sync to a queue a person sees. API keys stay in the client's name and a handover document ships with the build.
Where records drift apart today.
The same customer exists three times
One spelling in the CRM, another in accounting, a third in the store. Nobody is sure which address is current.
Orders get retyped into the books
The order is right in the store and wrong in accounting. You hear about it when the customer disputes the invoice.
A sync fails and nobody is told
The connector stopped on a Tuesday. You find out the next week, when a customer asks where the order is.
What I build
Connections between the systems you already run, with written rules for which one wins and a queue for anything that fails.
- One owner per field
- For each piece of data, one system is the source: the CRM for contacts, accounting for balances, the warehouse for stock. The others follow it.
- Two-way sync where it makes sense
- A phone number edited in either place updates the other. A paid invoice flows one way only. I write the rules down and you approve them.
- A visible error queue
- Every failed sync is listed with the record, the reason and a fix suggested by AI. A person approves the fix, then it retries.
- Matching before creating
- Records are matched on email, tax ID or account number before a new one is made, so one customer stays one customer.
- Connections to older systems
- When a system has no API, I use its exports, scheduled files or the database under it, and I say plainly what that limits.
- A change log
- Which system changed which field, and when. If two records disagree, you can see how they got that way.
- HubSpot
- Salesforce
- QuickBooks
- Xero
- NetSuite
- Shopify
- ShipStation
- Calendly
- Zapier
Also part of this work.
- CRM setup and cleanup
- HubSpot, Salesforce or Pipedrive set up around how you sell: duplicates merged, every lead source recorded, and each conversation attached to the right client.
- Reporting across systems
- One report that reads from the store, the books and the CRM, where every number can be traced back to the record it came from.
- monday.com boards
- Boards connected to the systems behind them, so a status changes when the order does.
Common questions
Can you connect a system that has no API?
Often, yes. I use exports, scheduled files or the database underneath. On the first call I tell you what that route can and cannot do before you commit to it.
Should we use Zapier or Make instead of a custom build?
For a simple one-way hand-off, yes, and I will set that up instead. Two-way sync, matching rules and a shared error queue get fragile in those tools, so that is where I build custom.
What happens when a sync fails?
It goes to the error queue with the record and the reason, and the person you name is notified. It retries after the fix. Nothing is skipped quietly.
Does AI change our records on its own?
No. The sync rules are fixed and written down. AI only reads a failed sync and suggests a fix, and a person you name approves it before any record changes.
What if a vendor changes its API?
Then that connection fails into the error queue, where it gets seen, instead of drifting. The API keys stay in your name and a handover doc ships with the build, so any developer can make the repair.
Three systems that had to agree
Multi-Channel Commerce
NYCMode
One inventory truth across Shopify and every marketplace it sells on.
Lead engineer / fractional CTO · multi-year
Read the case studyFractional CTO · E-commerce
Tres Colori
The storefront you see, and the order-ops machine behind it.
Fractional CTO
Read the case studyLive-Commerce Platform
Goodbye Inventory
Built for resellers who refuse to compromise.
Fractional CTO / Lead Engineer · ongoing
Read the case studyTell me which two systems disagree. I'll map the sync.
Tell me what you run and roughly when. If it's a fit, I reply within one business day to set up a free 15-minute call, then send a fixed price for the first paid step. If it isn't, I'll point you somewhere better.
New York, NY · remote-first across the US and Israel · English and Hebrew