Data orchestration examples in action: inside Rippling's AI-powered marketing engine

What do AI-written emails, a dog named Augie, and a fully autonomous pipeline for uncovering hidden markets have in common?

They're ALL part of the secret sauce behind Rippling's next-level marketing ops engine.

In a recent AMA-style webinar, Caden Bergeron, senior solutions engineer at Openprise sat down with Conrad Millen, VP of growth operations at Rippling, to unpack how Conrad's team is using automation, machine learning, and the Openprise RevOps Data Automation Cloud to run a highly personalized, always-on outreach engine at scale.

Sound familiar? Conrad covered this same topic live at MOps-Apalooza, but this time, he dives into the nitty-gritty details.

Whether you're trying to scale your outbound efforts, break into new markets, or just trying to get your hands on cleaner data, this conversation was packed with real-world tactics.

Let's dive in.

Why Conrad keeps coming back

Conrad's no stranger to Openprise—he's a three-time customer across his career at Okta, Clari, and now Rippling. And each time, he's found new ways to use the Openprise platform:

• At Okta: Openprise helped merge and clean data across two Salesforce instances post-acquisition.
• At Clari: It powered multi-touch attribution by aggregating data and feeding it into Salesforce.
• At Rippling: It's the engine behind autonomous, always-on marketing programs—and we're talking next-level, AI-personalized, don't-blink-or-you'll-miss-it scale.

I'm a three times Openprise customer now, and each business that I've brought Openprise into, we've used it in different ways, which is a testament to the flexibility of the product.

–Conrad Millen

Scaling personalized outreach with AI

Rippling's outbound engine, internally known as "me outreach" as in salespeople writing emails "themselves" (wink wink), is designed to send emails on behalf of sales reps at scale, but with a key twist: each message is uniquely written and highly personalized.

The ultimate goal wasn't to replace a sales rep. It was to make them more efficient. A lot of the manual research and the things that they would do to write that same email. We attempted to basically automate on their behalf and have them spend more time doing high impact activities like calls and outreach.

–Conrad Millen

Here's the breakdown of how AI is applied to Rippling’s emails:

• 50% of emails: Written entirely by AI.
• 30%: Tokenized personalization using vendor data ("Hi {{First Name}} at {{Company}}"… you know the drill)
• 20%: Human-written templates with minimal personalization.

The AI-generated emails aim to go beyond surface-level personalization. For example, one message references a prospect's recent LinkedIn post about their dog Augie while also tying in a relevant Rippling feature around tax registrations across states—demonstrating personal awareness AND business value.

Examples of Rippling’s data orchestration

Before we get into the architecture, it's worth naming what's actually happening here: this entire system is a masterclass in data orchestration — the automated coordination of data movement, transformation, enrichment, and activation across multiple systems.

If you've ever wondered what data orchestration looks like in a real production environment, here are five concrete examples from Rippling's own stack:

• Domain aggregation: Pulling net-new company domains from review sites, analytics tools, and third-party enrichment vendors into a single pipeline — no manual intervention required.
• Multi-vendor enrichment waterfall: Running each new domain through a sequence of enrichment APIs (the "Lead Search Waterfall") to populate contact personas, firmographic data, and fit scores — using the next vendor in the chain only when the previous one returns incomplete data.
• Intent signal aggregation: Every click, visit, review, and interaction gets captured into a custom Salesforce object, giving sales reps a real-time, unified view of account engagement across all touchpoints.
• Predictive scoring and automated routing: A machine learning model scores accounts and automatically routes them to the right rep — no spreadsheet, no Slack message, no manual triage.
• AI-triggered outreach enrollment: Once routed, contacts are automatically enrolled in AI-personalized email sequences — the 50/30/20 model described above — without a human ever touching the record.

Each of these is a discrete data orchestration example that most GTM teams still do manually, in parts, across disconnected tools. Rippling stitched all five into a single always-on engine. That's what separates a great data orchestration strategy from a fragmented one.

The Dark TAM engine: marketing ops wizardry

The standout moment of the webinar was Conrad's Dark TAM engine, an always-on, fully automated pipeline that identifies new domains, enriches contacts, routes leads, and scores accounts—no energy drink required.

Here's how it works:

1. Aggregate net-new domains from review sites, analytics, enrichment vendors, and more.
2. Validate websites, apply fit criteria, and create accounts in Salesforce.
3. Layer on "intent and event signals"—every click, visit, review, or interaction goes into a custom object, giving sales reps visibility.
4. Run the Lead Search Waterfall (shoutout to LSW), pulling in the right personas via third-party APIs.
5. Score accounts with a predictive model, route them, convert leads to contacts, and enroll them in AI-personalized outreach sequences.

A lot of this is powered by…Openprise.

Openprise is sort of the engine that helps us execute a lot of the data tasks here.

What makes it all work: speed, scale, and smart automation

When Caden asked what Openprise automation Conrad's team could live without, there was no hesitation.

A few standouts:

• API-enabled bots — gives him the ability to execute a series of jobs from an API call
• Infer task templates, which Conrad called "basically a SQL join at scale" — to run against other jobs and data sources
• And the unsung hero: data audit logs

Auditing for sure. Huge. Couldn't live without it.

–Conrad

Of course, smart automation means nothing without the right metrics. Rippling doesn't just stop at open or click rates—they use a machine learning model to analyze sentiment replies and track what really matters: the interest.

The key metric for email performance is how many positive replies you can generate.

But the best part? Rippling brought this entire system to life in just 6-7 weeks. A true testament to a team that knows how to move fast and build with confidence.

From data orchestration to AI orchestration: what's next for teams like Rippling

What Conrad built at Rippling is a data orchestration architecture. What teams like his are moving toward next is AI orchestration — where AI agents don't just write the email at the end of the pipeline, they make decisions throughout it.

Openprise's AI Agent Factory is built specifically for this evolution. Rather than bolting AI onto an existing workflow as a last step, it lets ops teams embed AI decision-making at any point in the data pipeline — scoring, routing, classifying, and enriching — with the same auditability and governance that Conrad called out as indispensable.

The outcomes are significant. As detailed in Openprise's AI orchestration white paper, organizations deploying AI within an orchestrated data pipeline — rather than as a standalone point solution — achieve dramatically better results than those layering AI on top of fragmented, unclean data:

• CrowdStrike used AI-powered behavior scoring within an orchestrated pipeline and achieved a 100% improvement in Inquiry to Open Opportunity conversion rate.
• Denodo automated persona classification across their contact database and successfully classified 90%+ of leads — resolving job title exceptions that older methods had missed for years.

The throughline is the same whether you're Conrad's team at Rippling or an enterprise ops team starting from scratch: AI is only as good as the data orchestration layer underneath it. The API bots, the infer templates, the audit logs — all of that is the foundation. The AI is what you build on top of it once the foundation is solid.

Want to go deeper on how AI orchestration works in practice? Read the full AI orchestration white paper.

Key takeaways

• Rippling uses Openprise to automate everything from account creation to AI-written emails — a real-world data orchestration example at full scale.
• The "me outreach" program is 50% AI, 30% token, 20% human — and all about scale with personalization.
• The Dark TAM Engine helps their sales team automatically find and engage new markets through five distinct data orchestration steps.
• Openprise helps stitch together data, orchestrate jobs, and keep things clean and compliant — and its AI Agent Factory is where teams like Rippling are heading next.
• New to data orchestration or AI for ops? Check out the Ops AI Prompts handbook and the AI orchestration white paper for practical next steps.

Want to scale outbound, break into new markets, and make your outreach smarter? Take a page from Rippling's playbook. With the right mix of automation, AI, and a powerful RevOps data automation platform like Openprise, you can build a personalized, always-on marketing engine that runs itself (almost).

Watch the full webinar or connect with us.