Every Vancouver business I talk to wants to know the same thing: can AI actually produce content that ranks, converts, and doesn't sound like a robot wrote it? The short answer is yes — but only if you build your AI content strategy around human judgment instead of trying to replace it.
I've spent the last 18 months building content systems for clients using Claude Code and other AI tools. Some experiments worked. Some failed badly. What I've learned is that the businesses getting real results aren't the ones publishing 100 AI-generated blog posts a month. They're the ones using AI to scale the parts of content production that don't require creativity — and then layering in editorial oversight where it actually matters.
Here's what an AI content strategy that actually works looks like in 2025, based on what I'm seeing with my own clients and the broader market.
The Three-Layer Model for AI Content Strategy
The best AI content strategies I've seen follow a three-layer model: automate the research and structure, use AI to generate first drafts, and always involve a human for the final edit and quality control.
Layer 1: Research and structure. This is where AI excels. I use Claude Code to pull keyword data, analyze competitor content, identify content gaps, and generate structured outlines. This used to take hours of manual work. Now it's a 15-minute process. The output is a detailed brief that tells you what to write and what angle to take — but it doesn't write the content yet.
Layer 2: First draft generation. Once you have a solid brief, you can use AI to generate a draft. The key is to treat it like a draft from a junior writer — it will have the structure and hit the main points, but it won't have voice, nuance, or the kind of insight that makes content worth reading. This is where most people stop, and it's why most AI content fails.
Layer 3: Human editing and refinement. The final layer is where you add the value that AI can't replicate: brand voice, expert perspective, real examples, and the kind of specificity that builds trust. This step usually takes 20–30% of the time a full manual draft would take, but it's what makes the content good enough to publish under your name.
If you skip layer three, you're publishing commodity content. If you skip layers one and two, you're leaving efficiency on the table. The strategy is doing all three.
Where AI Content Strategy Works Best
Not every content type is a good fit for AI. Here's where I've seen it work consistently:
- Product descriptions — especially if you have hundreds of SKUs and need consistent tone and structure
- FAQ content — answering common customer questions in a structured, searchable format
- How-to guides — step-by-step content that follows a clear process
- Local service pages — creating landing pages for different cities or service areas without rewriting the same content 50 times
- Content refresh — updating old blog posts with new data, examples, and structure
Where it doesn't work as well: thought leadership, opinion pieces, case studies (unless you're just using AI for structure), and anything that requires deep subject matter expertise or original research.
The pattern I've noticed: AI content strategy works best when the task has a clear structure and known best practices. The more the content depends on unique insight or subjective judgment, the more human involvement you need.
How to Build Your Content Workflow
If you're starting from scratch, here's the workflow I recommend to clients:
- Start with keyword research. Use your preferred SEO tool to identify target keywords. Feed the list into Claude Code and ask it to classify search intent (informational, commercial, navigational) and suggest content types for each keyword.
- Generate content briefs. For each keyword, create a structured brief that includes: target word count, suggested H2/H3 outline, related keywords to include, competitor content angles, and internal linking opportunities. I covered this process in detail in my post on automating SEO with Claude Code.
- Produce first drafts. Use AI to generate drafts based on the briefs. Set clear constraints: tone, reading level, keyword density, formatting rules. The better your brief, the better the draft.
- Edit for voice and accuracy. A human editor reviews each draft, adds examples, tightens the language, and ensures it matches your brand voice. This is also where you catch factual errors or generic statements that need to be replaced with something specific.
- Publish and track performance. Use your normal content calendar and analytics. Track which pieces rank, which convert, and which underperform. Feed that data back into your brief templates.
For most businesses, this workflow cuts content production time by 50–60% without sacrificing quality. The time you save on drafting gets reinvested in editing and promotion.
The Quality Control Problem
The biggest risk with any AI content strategy is publishing bad content at scale. Google's algorithms have gotten much better at identifying low-quality AI content, and the manual review teams are explicitly looking for it. If you flood your site with thin, generic content, you will get penalized.
The businesses that get this right treat AI as a drafting tool, not a publishing tool. Every piece of content that goes live should be reviewed by someone who knows the subject and can vouch for the accuracy and usefulness of what's being published.
I tell clients to apply the "would I link to this?" test. If you wouldn't feel comfortable linking to the piece from another high-quality article, it's not ready to publish. AI can help you produce more content, but it can't decide what's worth publishing — that's still a human call.
AI Content Strategy for Local Businesses
If you're running a local business in Vancouver or another city, AI content strategy has a specific use case worth calling out: creating location-specific landing pages and service content.
A lot of local businesses need the same core content repeated across multiple service areas. A plumber needs separate pages for "plumber in Kitsilano," "plumber in Yaletown," "plumber in Burnaby," and so on. Writing these manually is tedious, and they often end up thin or duplicate.
AI can generate these pages at scale — but only if you give it enough unique input for each location. The brief should include local landmarks, neighborhood-specific needs, service area boundaries, and local competitor context. The output should feel like it was written by someone who knows the area, not a template with the city name swapped in.
I've written more about this in my guide to how Vancouver businesses are using AI, but the short version is: local content needs local knowledge, and AI needs you to supply that knowledge in the brief.
The Tools I Actually Use
People always ask what tools I'm using. Here's the stack that works for me and most of my clients:
- Claude Code for content briefs, outlines, and structured research
- ChatGPT (GPT-4) for drafting when I need a more conversational tone
- Jasper or Copy.ai if a client is already using one and wants to stay in their ecosystem
- Clearscope or Surfer SEO for keyword optimization and content scoring
- Grammarly for final copy editing
The specific tool matters less than the workflow. I've seen people get great results with ChatGPT alone, and I've seen people waste money on expensive tools they don't use properly. Start simple, nail the process, then add tools as needed.
What to Avoid
Here are the mistakes I see most often when businesses try to implement an AI content strategy:
- Publishing AI content without editing. The fastest way to tank your site's credibility and SEO is to publish unedited AI drafts. Always review and refine.
- Chasing volume over quality. Publishing 50 mediocre posts won't move the needle. Ten well-researched, edited posts will.
- Ignoring E-E-A-T. Google's quality guidelines emphasize experience, expertise, authoritativeness, and trust. AI can help you produce content, but it can't make you an expert. Your content needs to demonstrate real knowledge.
- Using the same prompts for everything. Different content types need different approaches. A product description prompt won't work for a thought leadership post.
The common thread in all of these is treating AI like a replacement for strategy instead of a tool within a strategy. The strategy is still yours to build.
Getting Started This Month
If you want to test this for your business, here's a simple 30-day plan:
- Week 1: Pick 5 keywords you want to rank for and generate content briefs using AI
- Week 2: Use AI to draft one post based on the best brief, then edit it yourself
- Week 3: Publish the post and track its performance (rankings, traffic, engagement)
- Week 4: Refine your brief template based on what worked, then scale to the remaining keywords
Start small, measure what works, and scale deliberately. That's how every successful AI content strategy I've seen was built.
If you want hands-on help setting this up for your business, I offer AI marketing automation consulting specifically for Vancouver companies. And if you're still figuring out whether AI is the right move for your content, the FAQ page covers most of the questions I get asked on discovery calls.
The tools exist. The question is whether you're ready to build a system around them.