Buffer/LangChain/Mintlify的多人协作AI工作流搭建与治理

👋 This is a monthly free edition of MKT1 Newsletter—a deep dive into a B2B startup marketing topic, brought to you by Framer, Profound, and Closing Media.

⬆️ Upgrade to a paid subscription for: Full access to our MKT1 MCP | 100+ templates & resources | Post to MKT1 Job Board | MKT1 Newsletter Archive | $40K+ in discounts in the MKT1 Perk Stack

📅 Just announced: MKT1 x Lovable Buildathon on 10/28 at 10AM PT. Come build apps and MCPs alongside Kramer and the Lovable team. RSVP →

Subscribe now

Agents that are more useful than interns. A clean library of skills without gaps or duplicates. A GitHub repo every marketer uses. AI that makes sales and marketing the best of friends. The ability to run your skills anywhere you want. Self-updating, self-maintaining, never-stale routines. Connected MCPs for every tool you use with no feature gaps...the stuff marketing team leaders' dreams are made of?!

We all want a multiplayer AI system that makes our team more efficient, and we’re all stumbling our way there. Right now there’s an overload of new tools, models, and “hacks” and not enough clarity on how to pull them all together into something efficient and effective for your team.

“We’re still in the early innings of what an AI-native GTM org looks like.”

–Danny Lambert, Head of GTM Engineering, LangChain

I described many of the attributes of an ideal multiplayer AI system in the first newsletter in this series last month. In my definition, it has 4 parts:

But I wanted to bring this to life and give you a real glimpse into what teams are doing here right now, instead of leaving you with frameworks only. So, I chatted with 3 team leaders from 3 different companies. I teased some of their quotes and thinking in the first newsletter, but this newsletter details their setups.

I selected these 3 very intentionally: They're all in different leadership roles, but they all own pieces (or all) of the AI system for marketing. Each of the companies has a unique head start or advantage for building a multiplayer system. And while I talk about this stuff constantly, these three conversations changed or expanded my thinking about building for multiplayer AI.

The variability in how these teams have approached this same problem is in itself a lesson: You can't just copy another marketing team’s setup. Take the best practices and adapt them to your specific company. This holds for building out multiplayer AI and across all of marketing. As I always say, find your advantages, and double down.

“The problem is an AI workflow without context is running at maybe 60% of its potential. Your business context is what makes the output relevant and genuinely useful.”

–Simon Heaton, from his LinkedIn post about our chat

Recommended products & agencies

We only include sponsors we’d recommend personally to our community.

Framer is the AI website builder that has agents inside of the canvas. You get the speed and capabilities of an agent working alongside you, while maintaining structure and control for a full company site.

🎁 Offer: Get 30% off annual pro via the code MKT1

—

Profound: Understand and control your brand’s presence in AI Search. Track how your brand shows up across LLMs, then have its background agent—Aim—surface and execute on high-impact opportunities.

🎁 Offer: Get a free AI Visibility Report and see how your brand is showing up.

—

Closing Media is a boutique agency that runs LinkedIn and Google campaigns for growth-stage startups. They know the best ways to build and scale ad campaigns in 2026 (hint: thought leader ads and video!) and have years of expertise (I worked with them in 2018!).

🎁 Offer: Reach out to Closing Media here & mention MKT1 to get $5,000 of free ad spend for your next campaign.

In this newsletter:

This is part 2 of 3 in our newsletter series about multiplayer AI for marketing teams.

This newsletter:

• How 3 real teams build and run multiplayer AI: Buffer, LangChain, Mintlify:
• How they handle each of the 4Cs in multiplayer AI framework
• What I’d steal from each of their setups
• Multiplayer AI FAQs: Who should own the system, if humans should stay in the loop, how to keep your system from going stale, and more.

Part 1 - In case you missed it: Marketing teams are stuck in single-player Claude mode. Here’s how to go multiplayer.

Part 3 - Coming later this month: A deep dive into team MKT1’s multiplayer AI system

And a heads up: MKT1 Newsletter paid subscription prices will increase for new subscribers on October 1, 2026. Subscribe now to an annual plan to lock in $99/year for 2 years. More details at the end of the newsletter.

Subscribe now

RSVP: MKT1 x Lovable Buildathon on Wed 10/28 ➜

10am PT / 1pm ET, 45 minutes

Come build in Lovable with us. Bring an idea, and 45 minutes later, leave with something useful.

We’re hosting a buildathon with Lovable, the easiest place to build apps and now MCPs that bring those apps to Claude, ChatGPT, etc.

• I’ll build two apps live:
• A brand guide app that checks your marketing assets against your guidelines
• A benchmark tool for your customers built from real report data
• Then you’ll start on your own build, with 20 prompts to pick from
• Fern and Skyler from the Lovable team will answer questions

RSVP

Buffer: Growth marketing manages the system, the whole team contributes

I talked to Simon Heaton, Senior Director of Growth Marketing and Data at Buffer. Before Buffer he spent 6.5 years at Shopify, going from content marketing to Senior Growth Lead, and he's been building Buffer's growth engine since 2022.

"That's just the culture: high agency, high ownership.

We're small, independently owned, and bootstrapped. We're pretty agile, there's no heavy procurement. We can build and ship and deprecate as needed. That gives the team a lot of agency and ownership to just do it."

–Simon Heaton, Senior Director of Growth Marketing and Data, Buffer

The marketing team at Buffer runs on what Simon calls their marketing OS, built with Claude Cowork and Code, with skills shared through a GitHub plugin and context stored in Notion.

Buffer's multiplayer system works because the process fits into the culture they already have. They have a small team that's been around for years, public salaries, an open revenue dashboard, a 4-day work week, and a unique financing structure (they bought out their investors!). This is all well-documented and pretty interesting if you want to go down a rabbit hole...

Given that culture, contributing to the AI system is just part of everyone's job at Buffer. The team wants to build stuff and share it, not keep it to themselves or work in silos. This is ideal for building multiplayer systems, they are trying things out and learning fast.

This example also stands out because I get a ton of questions from people about ownership for multiplayer AI and AI GTM tools generally, and Buffer has a really solid setup. It achieves 2 things: Everyone becomes a builder (and gains a Gen Marketer skillset), and there's still a keeper of the overall system with clear ownership for how it all works together. It's both decentralized and centralized at the same time!

Here are the details of their multiplayer system:

Claude?

• The team works in Claude Cowork and Claude Code, whatever they feel most comfortable in.
• Skills reach everyone’s Claude through a Github plugin: One push to the shared GitHub repo and every team member's Claude picks up the update.
• MCPs connect Claude to the rest of their stack: Notion for context, plus tools like Profound and Granola that are called via individual skills.
• For more on how to use MCPs, read my last newsletter with 24 real MCP Workflows and a video recap from our MCP Showcase.

Context