Kleiner Perkins’ CFO used to spend up to six hours building a fully burdened cost model. Rippling AI did it in minutes.

The Challenge

Sue Biglieri is the CFO and COO at Kleiner Perkins, one of the most storied venture capital firms in Silicon Valley. In that role, she's responsible for decisions that require precise financial data — and every year, one of the most demanding exercises on her calendar is the annual compensation review.

The review itself is straightforward in concept: assess salaries, bonus eligibility, and the fully burdened cost of every employee. Fully burdened cost — what an employee actually costs the firm when you account for salary, employer payroll taxes, benefits, and 401k match — is a number that goes directly into the annual budget. It's cash out the door, and it's often more than people expect.

The problem wasn't knowing what the number should include. The problem was where all the pieces lived. Benefits costs varied by employee depending on dependent coverage and plan selection, so Sue would work from invoices. The 401k match required going to the 401k provider. Payroll gave her year-to-date figures but didn’t project forward or factor in employer-specific taxes like FUTA and SDI — items that were easy to overlook and difficult to calculate manually. And for bonuses, she'd go back to individual offer letters to determine eligibility and amounts.

Four sources, stitched together by hand, into spreadsheets that took five to six hours to build — with the persistent risk of an error somewhere in the process.

The Solution

When annual review season arrived this year, Sue tried a different approach. She opened Rippling AI and typed a single prompt:

“Can you give me the fully burdened cost of an employee for the whole year for 2026, by employee, by department?”

What came back impressed her immediately. Rippling AI started at the top — a single total figure for the whole firm — and then layered in progressively more detail. It broke out costs by category: salary, benefits, payroll taxes. It organized everything by department, pulling from the org structure already in Rippling without any manual setup. And it went all the way down to the individual employee level, so Sue could inspect the numbers for any single person on the team.

The departments were already preloaded in Rippling. The benefits data was there. The payroll figures were there. All of it came together in one response, within minutes, without Sue touching a spreadsheet or visiting a single external system.

Her first prompt hadn't included bonuses. But when she saw how accurately Rippling AI had handled the salary and benefits side, she realized it could do more — it had access to offer letters, and it knew what each employee was eligible for. So she asked it to add bonuses to the model. It did.

The output also included something she hadn't anticipated: key observations and assumptions, surfaced automatically alongside the data. Not just the numbers, but context about what the numbers meant.

The Impact

What used to take up to six hours of spreadsheet work was done in minutes. But Sue is quick to name the more important shift: what that time reclaimed actually made possible.

When the numbers take hours to build, the work of building them crowds out the work of understanding them. By the time the spreadsheet is done, there's little energy left to ask what it all means. With Rippling AI handling the construction, Sue had time to actually think — to look at what the firm is spending, by department, and ask whether it's going to the right places.

There was also the question of accuracy. Manual calculations of employer taxes — FUTA, SDI, and similar items — are easy to get wrong, especially when those calculations aren't a core in-house competency. In the past, some of those figures would be estimated. Now, Rippling AI calculates them from the source data, and Sue can verify specific entries against the underlying records.

She spot-checked the results. The numbers were right.

Six hours reduced to minutes. A fully burdened cost model spanning salary, benefits, employer taxes, and bonuses — across every employee and department — was produced from a single prompt.

One source instead of four. Benefits, payroll, taxes, and bonus eligibility — all pulled from data already inside Rippling, with no manual aggregation across external systems or offer letter lookups.

Accuracy, not estimates. Employer tax calculations that were previously estimated or prone to error are now computed directly from source data — with results Sue could verify line by line.

Time to lead, not just calculate. With the number-building automated, Sue had the space to do what a CFO is actually hired for: interpreting the data, comparing resource allocation across departments, and making better decisions.

For Sue, the security dimension mattered too. Running payroll data through an external AI tool wasn't an option she'd consider — the sensitivity of the information is too high. Rippling AI operates entirely within the Rippling platform, on data Kleiner Perkins already owns, which meant she didn't have to make that trade-off.

The annual review is now a different kind of exercise: less assembly, more analysis.