How We’re Building an AI-Native Sales Team at a 44-Person Company

Last Friday, I sat in a room with our sales team and said something that made a few people uncomfortable.

“If you can’t explain what MCP is to a prospect in two minutes, you’re already behind.”

Not because MCP is our product. Because MCP is what our buyers are asking about. And if the rep selling an AI-native platform can’t speak the language of AI natively, the gap between what we claim and what we demonstrate becomes the objection.

So we did something unusual. We stopped selling for a week and started learning.

The Problem Nobody Talks About

Every sales enablement blog tells you to “train your reps on the product.” Features. Benefits. Competitive positioning. We’ve done all of that. But there’s a different gap that training decks can’t close.

Our buyers are CTOs, VPs of Engineering, Revenue Operations leaders. They’re evaluating AI infrastructure. They’ve built custom GPTs. They’re experimenting with agents. They’ve connected their CRM to Claude through MCP servers. They speak this language because they live in it.

Our reps knew our product cold. But when a CTO said “We’re building our own agents — why do we need you?” the answer couldn’t be a feature comparison. It had to come from a place of genuine understanding. From having built an agent themselves and knowing where it breaks.

That gap — between product knowledge and ecosystem fluency — is where credibility lives or dies.

The Experiment: 6 Missions in 2 Weeks

We designed six experiential missions. Not training modules. Not videos to watch. Hands-on tasks that force our reps to USE the tools their buyers use.

Mission 1: Claude Deep Dive. Create a Claude Project for your sales workflow. Add custom instructions. Use it on a real deal. Experience persistent context. Then explain why it’s different from ChatGPT.

Mission 2: ChatGPT Power User. Build a Custom GPT. Not a toy — a real use case. Call prep assistant, objection handler, account researcher. Then understand why Custom GPTs are static prompts and GTM Buddy’s AI is contextual.

Mission 3: NotebookLM Explorer. Upload ten documents about a target account. Have a conversation with your sources. Generate an audio overview. Then know the answer when a buyer says “NotebookLM is free — why would I pay?”

Mission 4: MCP Decoder. Learn what MCP is. Watch GTM Buddy’s MCP server pull deal context into Claude in real-time. Then be ready when the CTO asks “What’s your MCP strategy?” This one broke the most brains.

Mission 5: AI Presentation Builder. Build a deck from a prompt in Gamma. Sixty seconds. Then understand why content creation is table stakes and the real question is who connects creation to deals.

Mission 6: Agentic Workflow Builder. The boss level. Build a multi-step workflow: Claude + GTM Buddy MCP + Slack. Research an account, pull deal context, draft a prep brief — one conversation. Then do the same task manually. Feel the delta.

Thirty minutes per mission. No reading. No tutorials. Just do it.

Why Experiential, Not Theoretical

There’s a 3-step sequence that governs how AI fluency actually works:

Step 1: Have you used it? Not watched a demo. Not read a blog post. Actually used the tool on a real task.

Step 2: Did you feel the power? The moment AI drafts a better email than you in ten seconds, or surfaces a deal insight you missed. That’s when conviction forms.

Step 3: Can you connect the dots? Take that personal experience and see it through the buyer’s eyes. Now you’re selling from truth, not from a slide deck.

Most AI training skips Steps 1 and 2 entirely. It starts at Step 3: “Here’s how to position our AI features.” But you can’t position what you haven’t experienced. The buyer knows the difference between a rep who’s used the tools and a rep who’s read about them. It’s the same difference between a travel writer who’s been to Tokyo and one who’s looked at photos.

What This Has to Do with Revenue Activation

We built Revenue Activation as a discipline. Five levers. Thirty diagnostic questions. A framework that identifies which part of your revenue system is capped and how to release it.

But here’s the thing nobody tells you about category creation: if your own team isn’t living the category, the market knows. You can’t sell Revenue Activation from an un-activated team. That’s not hypocrisy — it’s physics. The buyer is evaluating your people as much as your product.

So the AI fluency missions aren’t separate from Revenue Activation. They’re the foundation. A rep who has personally experienced what activation feels like — the moment AI surfaces exactly what they need, in the flow of work, without searching — can describe it with conviction because they’ve felt it. That conviction is the difference between “our product does this” and “let me show you what this feels like.”

What Happens Next

Session 2 is next week. The reps bring back their mission reports. We compare the manual approach to the AI approach for the same task. We name the delta. Then we connect every delta to a lever.

I’ll share what happens. The wins, the struggles, the things that surprised us. Building in public means being accountable to what you preach.

If you’re leading a sales team and wondering how to make them AI-native, here’s the honest truth: it starts with you. In the room. On a Friday. Doing the work alongside them. Not delegating it. Not buying a training platform. Living it.

“You can’t sell the future if you’re stuck in the present.”