The first time the Supermetrics team ran Flux AI Agent, the reaction was immediate.

"The first time that we used it, I think everyone was really impressed by how fast it was and how precise the explanations were," says a Financial Controller at Supermetrics.

That reaction is one we hear a lot. But for Supermetrics, the marketing data pipeline company trusted by over 750,000 businesses worldwide, speed and precision in financial reporting aren't nice-to-haves. They're the baseline for a finance function that has to keep up with a business moving as fast as theirs.

The problem with traditional flux analysis

Ask any financial controller what month-end actually looks like, and you'll hear the same story. Hours spent exporting data from the ERP, reformatting it in Excel, chasing down the transactions behind a variance. Then writing up explanations that someone on the leadership team will read once and move on from.

It's detective work, and it doesn't scale. Worse, it keeps finance teams from doing the work they're actually there to do.

For Supermetrics' accounting team, this friction had become routine. Data had to be pulled and converted before anyone could even start reviewing. Tracing a variance meant digging through the ERP by hand. Then the explanations had to be written from scratch, sometimes more than once a month.

Flux AI Agent: from raw data to explanation in seconds

Stacks' Flux AI Agent changes where the work begins. Instead of building reports before analysis can start, variances get surfaced directly, flagged automatically against thresholds the team configures, with context pulled straight from the ERP.

"Instead of digging into everything manually and converting the data first, we get the variances directly. The AI generation helps a lot."

That shift, from data wrangling to actual analysis, is where most of the time savings live.

Supermetrics now runs 17 Flux reports a month, with automated explanations covering 12% of their GL accounts. The explanations aren't summaries of numbers. They're narratives tied to the actual transactions and dimensions driving each change: which department, which vendor, which cost centre.

And when a variance needs a closer look in the ERP, no one has to go hunting for it. "You have the link already there, so you can directly look for it in the ERP system," the Financial Controller explains. "It's much easier and faster to review everything."

Quality that holds up to scrutiny

Speed only matters if the output holds up. Flux AI Agent generates audit-ready explanations, but the team stays firmly in the driver's seat.

The way Supermetrics uses it reflects how the tool was designed to work. The AI handles the heavy lifting on the first pass, and the team reviews, refines, and adds the context they have.

"We usually read through the explanations, and if there's something we know has happened, for example a bigger event with higher costs on a certain account, and the AI didn't catch that, we add it manually. But the explanations have been good so far, and I didn't need to add too much manually."

"It's a tool that gets the month-end review done efficiently, without compromising on quality."

What faster flux analysis unlocks

The real benefit of a faster, more automated flux process isn't just a quicker close. It's where the team's attention goes instead.

"When we don't spend that much time gathering data and doing the manual review, we can close a month faster and focus on more valuable and complex tasks."

That's the promise of agentic finance operations. Not fewer accountants, but more impactful ones. Time that used to disappear into data prep now goes into interpretation, judgment, and the kind of forward-looking analysis that actually moves the business forward.

For Supermetrics, this isn't a future state. It's how month-end works now.

Want to see how Flux AI Agent can work for your team? Book a demo.