At Abnormal, being an AI-native company means more than building cutting-edge security products. It means rethinking how we work, operate, and grow using the same technology that powers our platform. Across our teams, employees are embracing AI both to stop sophisticated threats for customers and to solve the company’s most complex internal challenges.

One of the clearest examples of this innovation comes from our AI Supervision team. The human-in-the-loop (HITL) group of 60 analysts operates 24/7 across multiple environments, managing 42 workflows that directly impact customer-facing queues and SLA performance.

When the team’s process for assigning analysts to workflows became an operational burden and a major cost in hours, the solution didn’t come from adding headcount or purchasing new software. It came from Brynn Collins, the team’s operations manager, who, despite having no technical background, leveraged AI to build an internal system that reimagined how the group coordinates its work.

For the AI Supervision team, staff coordination became one of the most persistent operational drains on productivity. Every hour brought the same challenge of reassembling the staffing puzzle so that analysts were aligned with the right workflows at the exact right moment.

The process was fully manual, and it unfolded the same way every hour: the coordinator needed to analyze queue volumes across all environments, send out Slack messages to track which team members were online and available, reference a spreadsheet to cross-match analyst skills and training status, then manually assemble an assignment list and send it back to the team.

On paper, the system worked fine. In practice, it consumed 15 minutes of every hour, seven days a week, adding up to more than 42 hours of labor weekly, the equivalent of a full-time employee. Worse, it was inherently reactive: staffing decisions lagged behind real-time conditions, and human error or bias could easily cause critical workflows to be under-resourced.

The team needed a solution that could match the pace and variability of their environment—something fast, flexible, and adaptive enough to keep coverage balanced in real time.

The answer came in the form of an automated coordination engine. The twist: it was not built by a team of engineers, but by an operations manager who knew the workflows inside and out.

The AI-powered engine transformed what had once been manual drudgery into a fully automated, intelligent system. It pulls real-time data from multiple sources: analyst availability through collaboration calendars, queue health through internal APIs, training status to ensure the right skill fit, and SLA risk across workflows. With these inputs, the system applies prioritization logic to match people with the work that matters most, then automatically posts an optimized staffing assignment list to team leads every hour.

What distinguishes this system from earlier attempts at automation is its ability to think beyond rigid rules. Previous efforts to solve coordination with spreadsheets or static formulas collapsed under the weight of real-world variability. Human schedules shift unexpectedly, workloads spike and recede without warning, and SLA thresholds move quickly. No static system could capture this dynamism.

By contrast, the new engine is built on AI prompting, iterative code generation, and adaptive logic, giving it the flexibility to respond to changing inputs in real time. The system doesn’t just follow instructions; it makes decisions that align with the way the team actually operates. It prioritizes critical queues first, ensures trained analysts are assigned appropriately, and continuously balances resources so production keeps flowing.

In essence, the coordination engine became a second brain for the team: fast, unbiased, and always on. What once required 42 hours of weekly human effort now runs automatically, with greater accuracy and consistency than ever before.

What makes this story extraordinary isn’t just the technology. It’s the human behind it.

The operations manager’s process began with a simple step: writing a detailed postmortem in ChatGPT Enterprise to explore potential solutions. From there, it became a collaborative back-and-forth.

The manager tested each iteration, spotting when the logic went off track. If code ran but produced the wrong result, she broke down why, shared errors, and clarified the outcome the system needed to deliver. Sometimes the AI’s output was technically correct but operationally wrong. Those were the moments where the manager’s deep understanding of the workflows mattered most. With each cycle, she refined the prompts, added context, and steered the AI closer to the logic that aligned with real-world conditions.

Over time, thousands of lines of code came together, not because of traditional programming expertise, but because of clear communication, persistence, and deep operational clarity. The breakthrough wasn’t just what AI could do—it was what a non-technical professional could achieve with AI as a collaborator.

Most importantly, success required understanding the problem with the depth of expertise that allowed the manager to recognize when AI wasn’t solving it correctly. That meant knowing what the right output should look like, anticipating edge cases, and proactively steering the system toward the intended logic.

This success underscores a larger truth: AI doesn’t replace human judgment, context, or creativity. It augments them. The human user provides vision, direction, and critical evaluation; AI delivers speed, adaptability, and execution. The magic lies in the partnership.

In just two months since inception, the coordination engine has delivered impressive results:

• SLA metrics have reached year-to-date highs across several high-impact workflows.

• Some workflows have seen SLA improvements of up to 40%.

• Customer-facing queues have improved by 7% quarter-over-quarter.

• Forty hours of manual work have been eliminated each week—equivalent to one full-time employee, with an annual ROI of over $45,000.

Just as important, the human impact has been profound. Instead of spending hours every week on staffing assignments, analysts can focus fully on production and customer outcomes. The system doesn’t just deliver better results; it creates a better experience of work.

This initiative reflects the AI-native philosophy built into Abnormal’s culture. Innovation isn’t limited to engineering or product; it is woven into every function. Employees across non-technical roles are equally trusted and empowered to experiment with AI, identify opportunities, and build solutions that drive internal impact.

For Brynn, the operations manager who built the system, the experience was transformative. What began as a coordination headache became an opportunity to discover new capabilities, both for AI and for herself. As she put it:

“I didn’t just discover what AI could do. I discovered all I could do with AI.”

This human element is what makes the story resonate beyond metrics. It shows that with the right culture and tools, anyone—not just developers—can innovate with AI.

The first iteration of the coordination engine is only the beginning. A second version is underway, incorporating learning capabilities and drawing on historical data and performance metrics to refine staffing decisions automatically. In parallel, the team is developing an Operational Forecasting Model that will predict workload spikes and staffing needs in advance.

These projects reflect a shift from reactive to proactive operations, combining the power of AI with the insight of the people who know their team’s needs best.

Abnormal’s mission has always been to harness AI to stop the world’s most sophisticated cyberattacks. But as this story shows, our commitment to being AI-forward extends far beyond our products. By empowering employees to use AI in their own daily workflows, we are redefining what’s possible inside the company as well as outside it.

The outcome is clear: greater efficiency, stronger performance, and a culture where anyone can leverage AI to transform the way work gets done.

To explore how the Abnormal approach to AI-native development powers our email security solutions, schedule a demo today.