Managing accounting complexity across 11 entities

Miro grew quickly through the shift to distributed work, expanding its operations across the US, Europe, Armenia, Japan, Singapore, and Australia. With that growth came new entities, currencies, bank accounts, and accounting processes.

Today, the team closes 11 entities across more than five countries, manages more than 25 bank accounts, and brings data from roughly a dozen systems into NetSuite. Madison Farnsworth, Head of Financial Systems at Miro, saw the complexity firsthand when she joined the company.

“As the business was scaling, we started hiring a lot in the US but also globally, and adding new entities around the globe. Today we have operations in the US, across Europe, in Armenia, and we’ve recently expanded into Japan, Singapore, and Australia. So a lot of complexity: different processes, different subsidiaries, foreign currencies across the globe.”

As Miro expanded, systems and processes were added to support new regions and workflows. Over time, teams in different regions developed their own ways of working, while accounting data moved between source systems, spreadsheets, NetSuite, and the close

Madison’s mandate was to standardize that operation and get Miro’s five-day close down to two.

Getting from a five-day close to two

Madison started by looking at every month-end process and asking why it was done that way. She found accounting data spread across systems that either weren’t integrated with NetSuite or still required manual exports into Excel.

Travel and expense was one of the clearest examples. Miro had adopted Brex and Navan as it grew, but the accounting team still had to recode card activity manually across 11 entities. Each month, accountants spent roughly 15 hours in spreadsheets applying coding rules, checking formulas and hard-coded overrides, reconciling the activity, and preparing it for NetSuite.

Cash journals, bank matching, and vacation and bonus accruals from Workday followed variations of the same process: export the data, reconcile it in Excel, build the journal entry, and upload it.  

Miro already used FloQast to manage the close. It gave the team a checklist, supported review, and provided auditors with evidence that reconciliations had been completed. The accounting work leading up to that checklist still happened elsewhere.

“It was a continuing handoff between different teams: get the export, get it reconciled, create the journal entry, get it into NetSuite, then pull it into FloQast to check it off. What we were looking for was all the steps that lead up to that checking off.” Those handoffs also made standardization harder. The US and EMEA teams had developed different close processes, creating more variation in how work was prepared, reviewed, and ultimately handed to auditors.

Evaluating whether to buy or build

Miro ran a deliberate evaluation of how to automate that work. The team looked at point solutions that handled individual parts of the close, including flux analysis and cash matching. Madison wanted to solve for the close as a whole, without introducing another set of handoffs between systems.

“Those were pieces of what we needed, but only a piece. Given the mandate to focus on it holistically and really tighten the close, where there are no more handoffs, the data is sent into one single place, and we’re auditing it end to end, the other tools just weren’t meeting that requirement.”

Miro also experimented with Claude and other general-purpose AI tools on close tasks. The pilots showed what was possible, but production accounting required repeatable outputs, controls, and an audit trail. “You need it to be much more deterministic and much more auditable. For processes that needed to be auditable at scale, those tools were not production ready.” Building internally meant that Madison’s lean team would be responsible for maintaining the pipelines, controls, and integrations as source data changed.

Replatforming onto an AI-native ERP carried a different set of costs and risks.

For Miro, Maxima offered a way to automate work around its existing ERP without taking on the maintenance and control requirements of building internally. Maxima already had SOC 2 in place and native integrations into the systems where Miro’s transactions lived.  “With a tool like Maxima, it’s a much easier ROI, because you can set it up without disrupting your core processes and the data in your core ERP.”

Taking T&E from 15 hours to 15 minutes

The Navan and Brex workflow became an early proof point. Maxima’s deployment team started with Miro’s Netherlands workbook, encoded the team’s existing accounting rules, and built an agent for each subsidiary. Each agent ingests card activity, applies the appropriate coding logic, flags transactions above defined thresholds or missing a GL code, and prepares the journal entry for review.

The workflow that previously took roughly 15 hours each month now runs in about 15 minutes. Accountants review the exceptions that require their attention rather than recoding transactions line by line.  

“That was the most time-saving one and the one that got everyone really excited. We set up agents, one per subsidiary, to automate that whole process by ingesting the data into Maxima and preparing the journal entry for us, which was the key piece. I wasn’t able to find that with any other vendor.”

Expanding automation across the close

• From T&E, Miro expanded into reconciliations, flux, cash journals, and accruals.

• Checklist management, account reconciliations, and flux moved into Maxima across all 11 entities, with multi-entity reconciliation workbooks and Slack-based approvals bringing teams into the same process. Max now prepares month-over-month and quarter-over-quarter flux explanations from the underlying transaction data.  

• The team applied the same approach to cash: JPMorgan and Citi activity across more than 25 accounts and multiple currencies is matched to NetSuite. Brex and Navan card activity, along with treasury intercompany and FX transfers. Hundreds of cash journal entries that treasury previously keyed by hand can now be prepared from the matched transactions for review.

• Vacation and bonus accruals are also moving out of downloaded Workday reports and pivot tables. Max prepares the entries while keeping employee-level detail out of NetSuite. Across these workflows, Miro preserves local-currency and USD views at the transaction level, along with the supporting evidence and reviewer behind each entry.  

Across each workflow, Max prepares the work and Miro’s accountants review the output and resolve exceptions.

Reliable productivity and compliance

As Head of Financial Systems, Madison also had to account for what happened after implementation: controls, auditability, changing source data, and ongoing maintenance. “It’s not just productivity. It’s reliable productivity and compliance.”

Miro needed controls it could test and evidence it could show its auditors. It also needed the integrations behind the automation to remain reliable as its financial infrastructure changed.

Bank feeds are a good example. Miro regularly adds accounts or encounters changes in bank-feed formats. Building internally would leave Madison’s team responsible for identifying those changes, updating the pipeline, testing it, and maintaining it alongside its other priorities.

With Maxima, Madison can work with the team to make those changes as they arise. “Any time we add a new bank account, or the format of our bank feed changes, which happens to us a lot, that’s something we can partner with Maxima on and set up very quickly. Versus me taking time out of my planned work, which is a disruption to the business, to pivot and do it ourselves.”

Scaling 2x without adding accounting headcount

Miro has doubled revenue without adding accounting headcount. Data that accountants previously pulled from separate systems and stitched together in spreadsheets is increasingly brought together and integrated with NetSuite through Maxima. The 15-hour monthly T&E process now runs in roughly 15 minutes, while US and EMEA teams work through the same governed close processes rather than maintaining separate ways of working.  

The close itself is moving toward the target Madison was given when she joined: two days instead of five. “The reason for a two-day close is that we can provide data for the rest of the business to make smarter, faster decisions,” she says.

There is still more work to bring into that model. Miro is looking at vendor timing and purchase-order accruals from Zip, commission accruals from Xactly, and continuous cash visibility across its bank accounts.

As those workflows move into Maxima, the goal is to remove more of the handoffs that slow the close today: pulling data between systems, preparing work in spreadsheets, and waiting for one team to finish before another can begin. Agents prepare more of the accounting work as the underlying data arrives, while Miro’s accountants review the output and focus on the exceptions and judgments that require their attention.