The Human Side of Automation: Knowing when to Step Inhero

AI and automation are changing everything about paid media, but not everything should be left on autopilot.

This session is built for PPC professionals who want to stay in control of their campaigns while letting machine learning do the heavy lifting.

Monica Valdez will show you how to read between the algorithmic lines, spot when performance needs a human course correction, and confidently refine automated campaigns across Google Ads and paid social without breaking what’s working.

Transcript

1. HEADSHOT HERE The Human Side of Automation: Knowing when to Step In Monica Valdez Green Blue Gray linkedin.com/in/mmvaldez/
2. "Most of the accounts I’m asked to fix aren't broken. They're misunderstood."
3. Your job has changed. THEN NOW Operator Interpreter Managing levers, adjusting bids, controlling every input. Reading signals, understanding context, knowing when to act.
4. What platforms do [mostly] well Bid optimization Budget pacing Pattern recognition at scale
5. What platforms still struggle with Context Business nuance Data quality
6. Where you still matter most Goal definition Setting the right target to begin with Signal prioritization Knowing which data actually matters Knowing when NOT to optimize Strategic restraint is a skill
7. The Signal Framework 3 signals to watch when automation isn't behaving the way you expect Volume Efficiency Input
8. Volume Signals What to watch: Conversions • Conversion Rate • Impression Share Warning signs: Flat volume despite budget increases Sudden volatility after learning completes Example: spend held steady. Calls dropped from 5-10 to 1 per week. The number that mattered wasn't on the dashboard.
9. Efficiency Signals What to watch: • CPA • ROAS Warning signs: Stable CPA with declining lead quality ROAS improvements driven by narrowing reach
10. Input Signals Conversion tracking setup Audience inputs Creative freshness Most Overlooked
11. If outputs look wrong, check inputs first.
12. Should you touch this campaign? Let's figure that out.
13. Step In When... Conversion volume drops but spend does not Lead quality shifts without structural changes Automation optimizes toward the wrong goal Creative fatigue is masked by algorithmic rotation
14. Step Back When... Learning phase is active Structural changes were just made Volume is trending positively with acceptable efficiency Fluctuations are short-term noise
15. Before you do anything: Pause. Observe. In that order. Act.
16. Take this with you. Step In vs. Step Back Decision Tree [QR CODE]
17. What happens when automation misreads the room entirely. It suggested "keyword" as a keyword. Maybe try adding “keywords” to your keywords?
18. Real Account. Real Decision. Account context Local services. Three campaigns, three deliberate budgets. What automation was doing Auto-applied portfolio bidding. One shared pool. Silent. Signal that was misaligned Blended CPA looked fine. Calls dropped from 5-10 to1 per week. Human adjustment Reverted. Deleted portfolio strategy. Disabled auto-apply. Outcome Restored in 5 days. Invisible change. Very visible damage. Focus: the decision process, not the platform.
19. You don't need to fight automation to stay relevant. Understand it. Guide it. Intervene strategically.
20. Automation executes. Humans decide. Connect with Monica