Voices of Search Podcast
Episode Summary: How to Set Up Gen AI Workflows While Future-Proofing for What Comes Next
Date: December 13, 2025
Host: Tyson Stockton
Guest: Eric Enge
Overview
This episode dives into the practicalities and challenges of integrating generative AI (Gen AI) tools into SEO workflows. The discussion focuses on how to strike a balance between operational efficiency and future-proofing—emphasizing the critical role of a “human in the loop” in AI-driven content and code creation. Tyson and Eric explore actionable strategies, pitfalls, and mindset shifts for building robust, adaptable SEO programs from the ground up.
Key Discussion Points & Insights
1. Starting an SEO Program with Gen AI (00:43 – 03:22)
-
Approaching Gen AI Integration
- Eric recommends leveraging tools like Claude or code-focused Gen AI but stresses that these are not flawless. Human review is always necessary to catch errors that AI might miss.
- “You still need human in the loop because the code is often flawed and sometimes in ways that could be hard to find too. So just be aware of that.” (Eric Enge, 01:08)
-
Areas Where Gen AI Can Aid
- Content ideation & production, organization of site content, schema writing, and even technical tasks (like .htaccess code).
- “There's so many different ways it can help… You just have to design your organization to have the right level of human in the loop.” (Eric Enge, 01:32)
-
Finding the Human/AI Balance
- The degree of human oversight should be tailored to your team's attitude toward AI—enthusiasts require more checks; skeptics, less.
- Start with “a bit too much human in the loop” and optimize as comfort with AI grows.
- “If you have the wild eyed visionary, then you probably need about 10 times more human in the loop than what they're thinking. And if you have the skeptic, you probably need half to a quarter of the humans that they they're thinking to find that average somehow.” (Eric Enge, 01:46)
-
The Imperative to Start Now
- Don’t wait for perfect clarity or boundaries—get in the game, iterate, and improve.
- A 30% efficiency increase is a realistic and transformative short-term goal.
- “Go for 30% increase in efficiency of your operations. You know how many people would have died for 30% five years ago? Oh my goodness, it's transformative.” (Eric Enge, 02:43)
2. Future-Proofing with a Human Touch (03:22 – 05:05)
-
Shifting Mindset During Traffic Declines
- Many are focused on negative traffic trends, but efficiency gains are a major “win” enabled by AI—even if traffic isn’t up.
- “There's other wins available that may not just be an absolute traffic… That efficiency gain and that potential for further output of activity from the team is such a key piece…” (Tyson Stockton, 03:31)
-
Why "Human in the Loop" is Crucial
- If anyone can prompt ChatGPT, what’s your value? Human curation and checks remain the differentiator.
- “The industry is littered with stories of people who actually started doing all their publications straight out of Chat GPT and they're site bombed…they tried to make too much of it.” (Eric Enge, 04:37)
3. Caution & Opportunity in AI Implementation (06:33 – 07:49)
-
Temper Ambition with Realism
- Don’t get swept up by hype—take incremental, “real” wins and keep your excitement/hormones in check.
- “You gotta keep your hormones under control. Hormones is probably the wrong word but your excitement under control and take the win, the real win.” (Eric Enge, 06:33)
-
Entrepreneurial Potential
- A technically adept, focused team could make substantial gains leveraging Gen AI, especially in niche spaces. But Eric, though tempted, admits he has “enough going on.”
- “There's probably a way to print money if you do that with the right small focused team… I've been dreaming about doing that for a while.” (Eric Enge, 06:46)
4. Application Suitability: Fault Tolerance in AI (07:49 – 08:49)
- Where Automation Can Fly Solo
- Some use-cases (like Meta’s AI-driven ad targeting) are “fault tolerant”—small errors aren’t disastrous—so they tolerate more automation with less human oversight.
- Not so for core website content or coding, where errors have outsized consequences.
- “If 10% of the ads are a really bad fit, it's not a crisis, the user just ignores the ad. So it's a fault tolerant application and those are situations where you can be less guided and get more scale. So that is not the content on your website…” (Eric Enge, 08:07)
Memorable Moments & Notable Quotes
- “[When] building an SEO program from scratch…you gotta use things like Claude code… as a fundamental part of how you develop stuff. Remembering that you still need human in the loop…” (Eric Enge, 01:02)
- “If you have the wild eyed visionary, then you probably need about 10 times more human in the loop than what they're thinking.” (Eric Enge, 01:46)
- “Just take the win that's on the table and don't try to make too much of it and have some fun along the way.” (Eric Enge, 02:59)
- “If I can just tell ChatGPT to do everything for me, so can a competitor. So… what can you do to make sure there's real add, extra value?” (Eric Enge, 04:29)
- “The industry is littered with stories of people who actually started doing all their publications straight out of Chat GPT and their site bombed.” (Eric Enge, 04:37)
- “Keep in mind there are certain applications that are naturally fault tolerant... That is not the content on your website or the way your website is coded…” (Eric Enge, 07:49)
Takeaways and Recommendations
- Start integrating AI tools, but never eliminate human oversight—especially in pivotal areas like content and code.
- Aim for tangible efficiency gains (e.g., 30%) as a clear, realistic goal.
- Continuously calibrate the human/AI collaboration as your team learns and as technology evolves.
- Identify aspects of your SEO workflow where automation presents little risk, to maximize gains safely.
- Stay grounded—balance excitement about AI’s potential with caution and a pragmatic approach to its limitations.
For more from Eric Enge, check his LinkedIn or visit pilotholding.com.
