Personalization in marketing is often talked about, but rarely delivered beyond surface-level fields like first name or company. For event marketing in particular, emails tend to restart conversations rather than continue them, ignoring where a customer is in the pipeline or how they’ve engaged before.

That gap is what Teddy Babanao, from the Revenue Marketing Automation team, set out to close. Historically, event emails at Abnormal were built using standardized templates. While these templates handled compliance, branding, and basic segmentation, they treated every email as an isolated message.

Each send failed to account for existing context: prior conversations, customer maturity, or relationship status. As a result, outreach felt generic, even when valuable signals were available. This represented a missed opportunity, especially at scale, where personalization directly impacts engagement and conversion.

Teddy built an AI agent that integrates directly into the existing marketing request workflow through Asana. When a request is submitted as a “personalized email,” the agent takes over downstream execution.

Once the audience is defined, the agent pulls in the relevant context, generates individualized email copy for each recipient, and provides real-time status updates as it runs. The process includes a deliberate human-in-the-loop step, where marketers review a sample of generated emails before approving the full send.

After approval, the agent pushes the finalized content into Marketo as a custom object, where it is formatted and delivered like any other automated campaign, complete with personalized subject lines, sender details, and event-specific calls to action.

What previously required days of manual copywriting and QA now runs end to end in a matter of hours.

The key innovation isn’t just automation; it’s continuity. Each email is generated as part of an ongoing customer conversation, not a standalone blast.

The agent can generate hundreds of personalized emails in a single run while preserving relevance and tone. By embedding human review at the right point, the workflow maintains trust and quality without slowing execution. Over time, the need for manual review can be reduced further as confidence in the system grows.

This approach allows revenue marketing to operate at a level of personalization that simply wasn’t feasible before.

From an efficiency standpoint, the agent saves at least two full days per campaign and can generate up to 500 personalized emails in roughly three hours. From a performance standpoint, the team expects a 15–25% lift in click-through rates, translating to an estimated $25M in incremental influenced pipeline.

Just as importantly, marketing teams gain visibility into what’s driving engagement and can tie personalization directly to measurable outcomes.

Next on the roadmap is confidence scoring, which will allow the agent to self-assess how strong each generated email is based on the quality and certainty of the underlying data. This will further reduce review burden and increase trust in automation.

The team also plans to extend the agent beyond event emails into triggered and nurture campaigns, with upcoming pilots planned for RSA and ThreatStream. This work represents a shift toward marketing automation that doesn’t just send faster, but communicates smarter, with context carried forward at every step.