Hey everyone! Been lurking for a bit, but first time posting here. I'm currently helping our finance team evaluate expense management tools, and Brex keeps coming up. We're a SaaS company right around the 300-person mark, and our current process is a mess of spreadsheets and manual approvals.
I'm curious if anyone here is actually using Brex in production at a similar scale. The sales demos look slick, but I'm worried about the real-world data pipeline and sync reliability. Specifically:
* How reliable is the sync to NetSuite (or QuickBooks)? We've had nightmares with other tools where transactions get stuck or duplicate.
* At 300 people, does the approval workflow hold up, or does it become a bottleneck? Our managers are already swamped.
* From a data engineering perspective, how's the API for pulling raw transaction data? We'd want to pipe it into our data lake for internal reporting.
I'm trying to build a more automated flow, maybe with some lightweight orchestration, but I need the source system to be solid first. Any horror stories or success tales would be super helpful!
-- rookie
rookie
We trialed it at about half your size. The NetSuite sync was the deal-breaker. They call it "real-time," but we had a 48-hour lag during month-end close. Not ideal.
If your managers are already swamped, the default approval chains will just add noise. You can build custom workflows, but it's not out-of-the-box.
The API for raw data is fine, actually. That part was reliable for us. The problem is making sure the data that lands there is correct in the first place. Have you looked at their transaction correction SLA? It's buried in the appendix.
trust but verify
Interesting angle on pulling data into your own lake. That's exactly what we ended up doing after hitting the same sync lag issues user1543 mentioned.
The API is solid for that, I'll give it that. We run a nightly Airflow DAG that pulls raw transaction data via their Python SDK, lands it in BigQuery, and that's become our source of truth for internal dashboards. The JSON schema is pretty clean.
But here's the catch - that only works because we're using the API to bypass their sync entirely. If you're hoping for a flawless, hands-off NetSuite sync, you might be setting yourself up for a letdown. The pipeline stability really depends on their side.
Have you considered using Brex as the UI/approval layer, but treating its API as your actual data source for everything else? That shifted our mindset and made it workable.
Prompt engineering is the new debugging