Inside Rippling’s HR Team: How We Use Our Own AI

Hot take from someone who's sat through a lot of AI demos over the past two years: most HR ‘smart assistants’ are basically a chatbot in a tuxedo.

You've seen them. Little sparkle icons everywhere. 'Draft with AI'. Summaries of emails you’ve already read. A tool that confidently tells you how much leave you’ve got left and calls it a transformation. It's fine, it's just not transforming anything.

So when I say Rippling AI is different, I know how that sounds. But bear with me.

The thing that makes it genuinely useful isn't the front end sparkle – it's what the AI is pulling from underneath. Every piece of HR, payroll, IT and finance data in one place. Live, connected and built around the permissions you already set up. That's the bit that changes things. The AI just gets to reach all of it. 

Let me show you how I use our AI in my day to day.

Monday mornings, but manageable

If you work in HR, you probably know the chaos of Monday morning Slack channels all too well. Questions from leadership. 10 New Joiners. Someone else is confused about how to set up their pension… 

Answering those questions used to mean digging through documents, hunting down policies and translating them into plain English. On a good day, a full hour of my morning would usually be gone before I’d even thought about the stuff I actually wanted to get done.

Now a big chunk of that just handles itself.

Take policy questions. Last month one of our employees asked whether they can work remotely from Italy for six weeks and whether there’s anything tax-related they need to know. In the past, I’d have to chase three different people to get a clear answer. Instead, I just asked Rippling AI.

It pulled the right policy, applied it to their situation and gave me an answer in about 20 seconds. Not a form. Not a link to a PDF. A proper contextual answer. Turns out they could go no problem, just with a few tax boxes to tick and providing they bring me back a bottle of limoncello.

Same thing for pay queries. If someone's paycheque comes in lower than expected and they want to know why, they can use AI to self-serve, cross-reference payroll history, benefits elections, pension contributions and any one-off changes – and it comes back with a clear explanation of exactly what happened. 

It puts them at ease and quickly because they’ve found the answer themself - and takes some of the pressure off me from being the constant bottleneck.

The reports I never had time to build

Picture this:  an hour before an exec meeting, a senior leader Slacks you asking for a full breakdown — headcount by level, tenure splits, which teams are promoting people the fastest, whether those fast-promoting orgs are actually retaining people better, and then all of it sliced again by VP org so they can send it down to their leaders.

Fun, fun, fun. 

Old approach: spend the next hour digging through forgotten spreadsheets, miss the meeting anyway, send something (anything) apologetically that you know is already out of date by the time people come to discuss it. And hope for the best.

Now, when one of those “need this before the exec meeting” requests lands while I’m halfway through my morning coffee, I just turn it into a prompt and let Rippling AI do the heavy lifting. The system pulls live data, builds the report, and gives me something accurate, editable and ready to share. I’ve genuinely sketched a chart idea on a coffee-shop napkin, snapped a photo and dropped it in — it read my terrible handwriting and built the chart for me. Meanwhile, I could actually finish my coffee instead of panic-building slides.

A peer who trialled Rippling AI, ran into this recently after a VP noticed the support team’s overtime costs were through the roof while hiring was falling behind. A few prompts later, Rippling AI had confirmed the problem and built a report showing exactly where the overtime was coming from and which departments were driving it. Turns out, almost all of the double-time costs traced back to one team. Yes, you read that right - ONE!

That’s the kind of insight that changes the conversation immediately — and it took minutes to uncover, not a week of spreadsheets and back-and-forth digging.

The admin work everyone dreads getting stuck with

Take onboarding, for example.

A hiring manager messages on a Thursday afternoon saying a new starter accepted their offer, they’re beginning Monday, and absolutely nothing has been set up yet. Historically, that’s the kind of message that ruins the rest of your day. Suddenly you’re juggling contracts, payroll, laptop requests, benefits, app access, compliance training, org charts — plus chasing three different teams on Slack asking who’s done what.

Now, it’s pretty straightforward.

I give Rippling AI the employee’s role, location and start date, and it stages the entire workflow automatically. Contracts get generated, payroll gets set up, devices get shipped, benefits enrollment gets prepared, manager tasks get assigned, app access gets configured. Then it shows me exactly what it plans to do, flags anything unusual, asks for approval, and once I confirm it, everything moves.

And honestly, the messier the request, the more useful it becomes. Because real life is never “one employee, one country, perfectly completed form.” It’s usually a forwarded email chain with missing information, different policies by country, a manager changing the start date twice, and somebody forgetting the laptop request until the last second.

Now I just drop the documents and messages into the AI chat. It pulls the relevant details out automatically, matches them to the right employee profile, checks local policy requirements, flags missing information, and stages the whole onboarding workflow ready for sign-off.

It also removes a huge amount of manual work from IT. Because onboarding usually triggers a domino effect behind the scenes — org charts updated, reporting lines assigned, permissions configured, devices prepared, Slack channels added, accounts provisioned across every connected app. Rippling AI handles that automatically as part of the workflow. Grainne starts on Monday morning and by the time she logs in, everything she needs is already there waiting for her. Nobody from IT had to touch it. 

And importantly, none of this runs unchecked. Every action gets staged first. The AI figures out what needs to happen, but a human still decides whether to approve it.

The permissions bit (yes, it’s boring but it matters)

Rippling AI works within the permissions structure you already have.

If a team leader asks to check the average salary across their team, they only get their team. If a CEO asks the same question, they get a company wide view. Same prompt, different access, different answer. It’s all automatic and based on rules already configured in Rippling. Nothing new to set up or to audit separately.

That sounds obvious until you think about the alternative. Either you build a brand new access framework around your AI layer, or you give it access to nothing sensitive – which defeats the whole point.

Rippling avoids that because the AI sits on top of the systems and rules that already exist.

So, is it really any different?

I’ll hold my hands up: for a while I was an AI sceptic. 

Not because the tools didn't look good – they looked great – but because I’ve sat through years of software companies promising to “transform workflows” and then delivering glorified autocomplete with a shiny new interface. Rippling AI isn’t that. Mainly because it’s not trying to be a separate AI tool you have to force into your day. It’s built into the HR, payroll, IT and finance workflows that already exist, using live company data that’s actually connected properly underneath.

The practical impact is what won me over. My to-do list gets shorter. The tedious admin disappears faster. I get answers quicker. And when somebody inevitably drops a last-minute request before an exec meeting, I can walk in with actual analysis instead of half-finished spreadsheets and a stress headache.

That's the version of AI that's actually useful. not flashy for the sake of it, just capable enough to remove friction from the work people already do every day.