Can you help me improve my agentic SEO workflow?
Hey everyone, I’m looking for a sanity check on an SEO blog workflow we built for a small niche ecommerce site.
Not selling anything here and intentionally not linking the site. I’m mainly trying to understand whether the workflow is sound, where it is risky, and what experienced SEO people would change before we scale it further.
The site is in a developer culture / tech apparel niche, so the content strategy has a commercial and editorial posts around programmer gifts, developer shirts, coding merch, hackathon shirts, non-cringe coding gifts, etc.
Here’s the workflow we used.
First, we did competitor and SERP research. We looked at similar ecommerce/blog sites, their sitemaps, their indexed blog topics, and their content angles. The main gap we found was that most apparel competitors have product/category pages, but not much useful developer-specific editorial content.
Then we used Reddit research to find actual questions and wording people use around developer gifts and programming merch. We were careful not to just copy threads. The idea was to use Reddit as intent research: what buyers are confused about, what developers call cringe, what jokes are overused, what gift buyers worry about, etc.
From that, we built a hub-and-spoke plan:
- Hub: programmer gifts
- Spokes: software engineer gifts, coding gifts for beginners, useful gifts for programmers, programmer shirts, funny developer shirts, hackathon team shirts, programming merch jokes to avoid, etc.
Each Reddit-discovered question was mapped to exactly one post so we wouldn’t create 5 articles fighting for the same intent.
For the first draft batch, we generated 10 SEO drafts with these rules:
- Non-question SEO title
- One H1
- At least 5 H2 sections
- Key takeaways
- FAQ section
- Internal links to related posts/products
- At least one authority source in the draft
- No generic AI phrases
- Brand voice should be developer-native, specific, and a bit dry, not “ultimate guide” slop
The drafts passed a basic QA check, but we are not treating them as publish-ready yet.
Before publishing, the plan is:
- Run a deeper SERP brief per keyword
- Expand posts to around 1,500-2,500 words where needed
- Add 3-5 real image/product placements
- Add more external authority links
- Add BlogPosting, FAQ, Breadcrumb, Author/Organization schema where appropriate
- Run SEO, GEO/LLM visibility, schema, and content quality checks
- Publish as Shopify drafts first
- Manually review before making anything live
- Add reciprocal internal links after publishing
We also worked on the site architecture around the blog:
- Created a proper blog hub
- Split the blog into “Developer Culture” and “Tech Personalities”
- Added editorial category pages instead of default Shopify lists
- Added breadcrumbs and JSON-LD
- Improved title/meta handling
- Made article titles the anchor text instead of generic “read more”
- Kept the Shopify theme mostly Liquid/CSS without heavy JS
My questions:
- Does this workflow sound like a reasonable hybrid agentic SEO process, or is it still too close to an AI content factory?
- Is using Reddit for intent discovery in this way okay, assuming we are not copying posts or comments?
- Would you publish the gift/developer merch cluster first and pause the long-tail entity pages until they are manually improved?
- Are FAQ sections/schema still worth including on ecommerce blog posts, or would you focus more on direct answers, tables, and internal links?
- What quality gate would you add before letting an agent-assisted draft become a live indexed page?
I’m especially interested in criticism. I’d rather fix the workflow now than scale something that looks good internally but is weak from an SEO/AI citation perspective. Also just to add, I'm so new to SEO, I'm doing this because I don't want to rely on meta ads thats why I'm try to explore SEO to help with marketing, I just combined all the YT videos and github repo that I gathered to create this workflow for my store.