Treasury teams have done the hard work. The bank connections are live, the TMS is configured, payment workflows are running. That investment took years and millions. But if your team is still spending 40+ hours a month on manual categorisation, static spreadsheet forecasts, and variance reports that take three days to compile — the problem was never the infrastructure. It's that nobody built the intelligence layer on top of it.

Your TMS was designed to move money. It was never designed to think. 

The real problem isn't connectivity. It's intelligence

Most enterprise treasury teams have already invested significantly in their connectivity layer. They have bank feeds flowing into a TMS or ERP. They have SWIFT connections, SFTP drops, MT940 and BAI2 files landing where they should. The plumbing works.

What's missing is the layer above it: the part that takes all that raw data and turns it into something you can actually act on. Transaction categorisation that doesn't rely on brittle rules. Forecasts that learn from your data instead of sitting in a static spreadsheet. Variance analysis that explains itself. Cash positioning that tells you where idle balances are sitting and what to do about them.

This is the intelligence gap. Treasury teams have done the hard work of getting data into their systems. But nobody is doing the hard work of making that data useful; at least not without 40+ hours of manual effort every month.

The case for building on top

Here's the mindset shift: instead of replacing the infrastructure you've already invested in, what if you added intelligence to it?

Treasury teams have spent years configuring their connectivity, their bank relationships, their payment workflows. That investment shouldn't be thrown away every time you want better forecasting or smarter categorisation. The foundation is sound. What's missing is the thinking layer on top.

The technology exists today to sit above your existing TMS or ERP and turn that raw data into actionable insight without touching your source systems, without ripping out what works, and without a 12-month implementation timeline. The question isn't whether to modernise. It's whether modernising means starting over, or building up.

The technical debt nobody talks about

Every rules-based categorisation engine is technical debt. Every manually maintained spreadsheet forecast is technical debt. Every static report that takes three days to compile is technical debt accumulating interest.

Treasury teams have been compensating for that architectural limitation with headcount and hours for a decade. And the irony is, another rip-and-replace doesn't eliminate that debt — it just resets the clock on a different platform.

The smarter play is to pay it down. Capture your team's institutional knowledge — every correction, every override, every judgment call — and codify it into something that compounds over time. Your best treasurer's expertise becomes organisational infrastructure. That knowledge never retires, never takes holiday, never leaves for a competitor.

What ‘building on’ actually looks like

Here’s what an intelligence layer means in practice.

Think about categorisation. Most treasury teams rely on rules-based engines — static logic that breaks every time a new vendor, entity, or currency flow enters the picture. An intelligence layer learns your company's financial vocabulary over time, context-aware rather than rule-dependent.

Or take forecasting. The default approach is a spreadsheet someone updates weekly, maybe monthly. An intelligent machine learning layer retrains on your actual transaction history,daily , and surfaces variance explanations in language and as reports your CFO can actually read. No more spending three days building the same report before every board meeting.

Cash positioning is another one. Most teams can tell you what happened yesterday. Very few can tell you where idle balances are sitting right now, across entities, and what to do about them. That's not a connectivity problem. That's an intelligence problem.

Start from where you are

The treasury teams getting ahead right now aren't the ones running the biggest transformation projects. They're the ones asking a different question: what can we do with what we already have?

The data is there. The connectivity is there. The investment has already been made. What's missing, and what the next generation of treasury technology is finally solving is the intelligence layer that makes all of it useful.

So before kicking off another 12-month migration project, it's worth asking: is the problem really the TMS? Or is it the intelligence layer that was never there to begin with?

If it's the latter, you don't need to start over. You need to build on what you already have.

How Palm helps

This is exactly why we built Palm. Not to replace your TMS, but to make it smarter.

Palm sits on top of your existing infrastructure as an AI-native intelligence layer. It connects to your banks, ERPs, and TMS through a read-only, zero-write architecture; meaning nothing is disrupted, no source systems are modified, and your CISO isn't losing sleep.

From there, Palm does what your current systems weren't designed to: it categorises transactions using context rather than static rules, retrains forecasts daily on your actual data, surfaces idle cash across entities, and delivers variance explanations in plain language. 

One example: when ON, the global sportswear brand, connected Palm to their existing treasury setup, they weren't looking to replace anything. They wanted to see what their data was already telling them. Within weeks, Palm had surfaced $350M in idle cash that was invisible in their existing reports. 

And that’s the kind of outcome that becomes possible when you build on what you already have.