The first live demo started with Corey Haines realizing he had forgotten the YouTube link.
A few minutes later, the transcript from that video had become a landing page and a newsletter. By the end of the hour, real keyword data had become a small programmatic SEO cluster, and a competitor page had received a second pass from Corey’s copywriting skill.
That progression captured the point of the session. AI can make content production much faster, but Corey kept returning to the same condition: the work has to get better before it gets bigger.
Corey is the CEO and Co-founder of Conversion Factory. After more than a decade in marketing, he turned his agency SOPs and experience into one of the largest open-source libraries of marketing skills. Those skills now run as Ploybooks.
The recording is below. The rest of this article walks through the 3 live workflows, the decisions Corey made inside each one, and the quality rules he refused to automate away.
Jump to a section
- SEO and AEO met in the same place
- Quality became the answer Corey kept returning to
- Then Corey opened the Ploybooks
- The first demo began with a missing link
- Programmatic SEO went from a 6-figure bet to a first cluster
- The last build moved into code
- The chat took over
SEO and AEO met in the same place
Search behavior has changed. Clicks and impressions are under pressure, and buyers are using more places to evaluate products. Corey has been watching the shift from the agency side. For a long time, SEO was a resource game. Large companies could spend on links, engineering, design, schema, and the site architecture needed to support it all. AI has made much of that work cheaper and more accessible.
For Corey, AEO is the next version of the same discipline.
“I really see AEO as essentially SEO 2.0, and I think they’re actually both kind of converging a little bit as well.”
“Good AEO is essentially great SEO,” Will added. The hygiene, legibility, and content that help a page perform in search also help answer engines understand and cite it. The labels are changing faster than the basic work.
That easier access has made the field more competitive too. If every team can produce pages, the advantage comes from what goes into them.
Quality became the answer Corey kept returning to
Corey had seen the same traffic graph more than once: a line climbs, plateaus, then drops when a team tries to turn AI into an automatic publishing machine. He remembered the early rush to generate 100 blog posts at once and the surprise when those posts did not perform. His answer was blunt. Take a few extra hours. Publish something genuinely good. Do not put out slop.
The strongest source material, he said, is usually already inside the company. It might be a meeting transcript, notes from a walk, a brain dump recorded on a phone, proprietary data, original research, a survey, or an interview. Those are the places where a team’s real voice, opinions, facts, and experience show up.
“The competitive edge now in AI, particularly when it comes to quality of content, is your data. What can you bring to the table that no one else can?”
The same idea shows up in engineering: before you automate a process, do it manually a few times. Learn what excellence looks like before you put the work on a loop. A bad starting point running 10 times faster only makes the team worse faster.
The rest of the hour followed that pattern. Each demo began with something the team had already made, then showed how a defined workflow could carry it further.
Then Corey opened the Ploybooks
Corey shared his screen and opened the Ploybooks built from his open-source marketing skills. He described a skill in the least mysterious way possible: a markdown file, a set of instructions, and any scripts, references, or templates the agent needs to do the work properly.
“A skill is just an executable SOP.”
A reusable workflow gives the agent Corey’s way of doing the job. Without those instructions, the model falls back to the same defaults everyone else is using.
The first demo began with a missing link
The first workflow was supposed to turn a video into a landing page and a newsletter. Corey clicked run, started the Ploy, and then realized he had forgotten the YouTube link.
He opened the public Conversion Factory channel, found the video, copied the URL, and came back to the workflow. There was another brief live-demo moment when the Gemini key needed attention, but the transcript came through. Ploy pulled out the themes, including the “spiky points of view” that sounded like Corey rather than generic marketing copy.
From there, the same source became a landing page and an email newsletter. Corey clicked through the finished artifacts in the browser while Will pointed out that the email could also move into the tools a team already uses.
Corey compared the process to pottery. The transcript was the raw material. The workflow shaped it step by step until the chaos had some order. A video happened to be the source in this demo, but it could have been meeting notes, an internal brainstorm, or an interview with a subject-matter expert.
You can run the same workflow with the video-to-landing-page and newsletter Ploybook.
Programmatic SEO went from a 6-figure bet to a first cluster
For the second demo, Corey opened his programmatic SEO Ploybook. A few minutes earlier, he had described what this work looked like 5 years ago: a marketer, a designer, an engineer, months spent scraping and structuring data, and a 6-figure investment that might take multiple quarters to pay off.
The Ploybook started with real keyword data from DataForSEO. It looked at the search results, worked through 12 programmatic SEO playbooks from Corey’s own experience, and recommended a direction. Corey changed the choice from a glossary to persona pages, and the strategy document adjusted with him. A first cluster of 5 pages took shape on screen.
Will asked what happens when a team gets this wrong. Corey had tried programmatic SEO the hard way, built teams around it, and even started a company around the problem before AI improved enough to change the economics. He had also watched companies flood their sites with low-quality pages and spend the next year trying to recover.
“Google doesn’t hate programmatic SEO. Google hates poor, low quality content.”
The first cluster stayed small on purpose. Corey suggested “one or 5 or 3,” any number a person could review manually, adjust, and learn from before asking the system for hundreds more. By the end of the demo, the plan had become a small set of connected pages rather than a promise to generate 1,000 URLs and hope.
Run the full method with the Programmatic SEO at Scale Ploybook.
The last build moved into code
With about 15 minutes left, Corey switched to a coding agent. The terminal looked nothing like the Ploy interface, but the way he worked was familiar. He asked for a quick plan and one competitor-alternative page for a fictional SaaS company. He wanted the first page on localhost before the system tried to scale it.
Competitor-alternative pages put your own facts and point of view into the source material when someone asks an answer engine about your product at the bottom of the funnel.
“You want to dictate the narrative. You want to feed it all the information that it needs.”
The first version of the page appeared. Then Corey asked the agent to run his copywriting skill across it. The structure stayed in place, while the wording shifted toward his rules for direct response, storytelling, and tone. One workflow built the page. Another gave the copy a second set of eyes.
The chat took over
The demos finished with enough time for the questions that had been collecting in the chat.
How many programmatic SEO pages are too many?
“No such thing,” Corey said, before adding the important condition: the number has to come from the keyword data and the strategy. Will suggested a simple way to think about it. If Google sampled 20 or 50 pages and handed them to a person, would that person say each page deserved to exist on its own? A useful marketplace might have hundreds of thousands of pages. Creating 40,000 pages without a reason is a different story.
When should you check a new page’s performance?
Corey joked that he checks the numbers “obsessively, like every 5 minutes.” In practice, he waits about 1 or 2 weeks. If a page still is not indexed, or Search Console shows a blocker, he fixes the issue and requests indexing again.
For AEO, he shared a rough test anyone can run: ask the same question before and after publishing the page and see whether the answer changes. It is not a replacement for prompt-tracking software, but it gives a team somewhere to start.
How do you build a copywriting skill that sounds like you?
Corey starts with material that already sounds like him. He suggested asking an agent to study your past chats, the words you use, and the way you phrase things. His own skills began with agency SOPs. Swipe files, user guides, and anything you would hand to a new hire can become source material too.
Can Ploy work alongside Framer or Webflow?
Will described the hybrid setups Ploy already supports. Some teams move the whole site. Others keep their existing homepage or community experience where it is and use Ploy for an SEO section, an ABM motion, or landing pages for paid campaigns.
We finished right on time
At 1 PM, Corey looked for another question and found an empty queue. Will pointed out how rare that is in the webinar world. Corey shared where people could find his marketing skills and offered to keep answering questions after the session. Will closed by previewing the next webinar, a session on AI-native ABM with the team at Avoca.
An hour earlier, SEO and AEO had felt like a changing set of rules. By the end, we had watched the same idea work 3 different ways. A strong source became more content. Real keyword data became a small page cluster. One comparison page got a second pass from Corey’s copywriting skill. In every case, the work got better before it got bigger.
Open Corey’s Ploybooks to try the workflows, or book a demo to map them to your site.