The Setup Most Operators Get Backwards
The default playbook: hire sales reps, then figure out the systems to support them. depthfirst flipped it. Mark Hardy, a solo RevOps hire, built the entire GTM architecture before a single account executive joined the team — and the coverage he achieved would have required a small department under the old model.
The case study published by Clay is the most concrete AI×RevOps example in recent memory. Specific stack, specific signals, specific workflow logic. Worth reading in full if you're designing or auditing a modern GTM system.
What the Architecture Actually Looks Like
The core insight is architectural, not tactical: make the signal layer do the coordination work that humans used to do manually.
Here's how Hardy wired it:
- Clay acts as the central signals repository, aggregating 150+ data providers — breach alerts, 10-K filings, intent signals — into a single layer that auto-updates downstream audiences without manual intervention.
- Gong call intelligence feeds back as first-party signals. Call outcomes don't stay in Gong — they become Clay segments that trigger campaigns automatically.
- Claude Code handles custom logic where off-the-shelf enrichment falls short.
- LinkedIn Ads audiences update dynamically based on signal combinations, not static lists someone refreshes on a schedule.
The result: a one-person function achieving precision targeting on security buyers using real-time data events, at a scale that previously required meaningful headcount to maintain.
The Takeaway for Operators
The bottleneck in most RevOps orgs isn't sales capacity — it's signal latency. Reps are working lists that are already stale. Campaigns are targeting audiences assembled last quarter. The manual coordination layer (someone updating a spreadsheet, someone pulling a report, someone uploading a CSV) introduces lag at every step.
What Hardy built eliminates that lag by making signals directly actionable. The system doesn't wait for a human to notice a trigger and route it — it routes itself.
Before your next sales hire, answer two questions: What signals should change how you target or message a prospect? And how many manual steps currently sit between that signal firing and a rep or campaign acting on it? If the answer to the second question is more than one, you have an architecture problem that headcount won't fix.
The systems-first sequence isn't just more efficient — it means the sales team you eventually hire ramps faster, works warmer lists, and doesn't spend cycles on coordination that the stack should handle automatically.