What's your brand's AI level?

Most brands are stuck at Level 2. Where does your brand sit?

October 21, 2025

And more importantly - where should your team be if they want to stay relevant?

Here's the progression I'm seeing across of e-commerce brands:

Level 1: Haven't Touched It

Still doing everything manually. Spreadsheets for inventory. Copy-pasting customer emails. Watching session recordings one by one.

You're not just behind - you're invisible to the AI agents that are increasingly making purchase decisions. As I wrote in "Humans will stop visiting websites", Jakob Nielsen predicts AI agents will replace humans browsing websites entirely. If you're at Level 1, you don't exist in that future.

Level 2: Used ChatGPT for Content

Writing product descriptions, email copy, or social posts with AI. You paste your brand guide into ChatGPT and hope for the best.

This is where most brands sit. It's better than nothing, but you're basically using a Ferrari to deliver pizza. And as I showed in "Agentic Brand DNA", copy-pasting brand guides into every prompt is killing your consistency and wasting time.

Level 3: Deployed Pre-Built Solutions

Added chatbots, AI writing tools, or basic automations with Zapier. You've got Jasper for copy, maybe Klaviyo's AI features, some basic flows.

You're automating tasks but not thinking. Your tools don't talk to each other. Every solution is its own silo.

Level 4: Built Custom GPTs/Assistants

Created brand-specific AI assistants using OpenAI GPTs or Claude Projects. Your team has a "Brand Voice GPT" that knows your guidelines.

This is progress, but you're still locked into one platform. What happens when your designer prefers Claude but your marketer uses ChatGPT? You rebuild everything twice.

Level 5: Connected AI Agents to Workflows

Built agents in visual workflow builder like n8n that actually do things. Maybe they process orders, handle support tickets, or generate content automatically.

Now we're talking. But if you're pulling data from CSVs and basic APIs, you're missing the real power. As I explained in "Why Your ChatGPT Data Analysis Keeps Failing", agents need proper database connections, not file uploads.

Level 6: Integrated MCP Tools as Clients

Connected your agents to existing MCP servers for file access, database queries, or API integrations. Your agents can now access real-time data.

This is where things get interesting. Your AI can query PostHog for user behavior (see "Agentic CRO via MCP Using PostHog") or pull Klaviyo data for retention analysis ("Agentic Retention with Klaviyo") or campaign memories for comparisons to past campaigns (“Agentic Brand Memory”)

Level 7: Built Custom MCP Servers

Developed your own MCP servers for inventory, Shopify data, or customer analytics. Your agents don't just read data - they access it through your custom interfaces.

Remember that "Agentic Commerce via MCP" demo where I built an MCP server for Finisterre? That's this level. You're creating proprietary data access that competitors can't replicate.

Level 8: Orchestrated Multi-Agent Systems

Multiple specialized agents working together using frameworks like LangGraph. Your pricing agent talks to inventory. Marketing agents coordinate with fulfillment. All connected through tools like LangGraph.

This is what we built with "Agentic Ad Making" - five different agents passing work between each other to create ads from raw footage. Or the influencer campaign system in "The Making of AI Agents for Influencer Marketing" where agents handle everything from discovery to payment.

Level 9: Achieved Autonomous Commerce

Agents with full MCP access making real-time decisions. Pricing adjusts automatically. Inventory reorders itself. Marketing spend optimizes continuously. Humans just set guardrails.

The "Agentic Shopify Admin via MCP" demo shows a glimpse of this - an agent creating products, running promotions, and managing customers through voice commands alone.

Level 10: Created AI-Native Infrastructure

Complete data liberation. Everything vectorized (remember "First Step to Agentic Brand: Vectorize"?). Custom MCP servers for every function. Agents as Slack coworkers handling daily operations.

Your "Scene Prep Agent" turning thousands of videos into searchable vectors? That's Level 10 thinking. So is having a "Graph Agent" that discovers hidden patterns in customer behavior that Shopify and Klaviyo can't see.

Level 11: Scaled AI Engineering Org-Wide

Every employee builds agents. Marketing creates campaign agents. Ops builds logistics agents. Support makes resolution agents. All interconnected.

Your creative team uses "Agentic Video Editing via MCP" right in Google Sheets. Your retention team has agents analyzing cohorts. Your CRO team has agents catching conversion issues before they cost money ("Agentic CRO with MCP: Anomaly Detection").

If you're below Level 5, it might be time to speed things up.

What You Need to Do Tomorrow

If you're at Levels 1-4: Start with one workflow. Pick your most repetitive task and build an agent for it. Could be "Agentic Compliance for Creator Content" if you work with influencers. Or "Agentic Testimonial Ads" if you need to evaluate videos.

If you're at Levels 5-7: It's time to build your data infrastructure. Vectorize your content. Build MCP servers. Give your agents real power. Start with "Your Brand Is Not Ready for AI Agents If..." to see what data needs liberating.

If you're at Levels 8-11: You get it. You're building the future. Now it's about scale and teaching your team.

Learn to think in systems. Understand how agents pass information. Know when to use vector search vs. graph databases vs. SQL. Build MCP servers that expose your unique business logic.

Because in 12 months, every brand will have agents.

Where's your brand on this scale? And more importantly - where's your personal skill level? Because if your company is at Level 2 but you can build Level 7 solutions, you just became the most valuable person in the building.

Want to level up? Start with one agent. One workflow. One problem solved. Then connect it to another. Before you know it, you'll be orchestrating systems that would have seemed like magic just last year.

If you’d like more guidance on upping your brand’s AI level, reply to this email or DM Bora on LinkedIn.