Week 3. Thank you for all the constructive feedback! Keep it coming. Bear with me as I'm iterating. You can simply reply to share your thoughts or workflows.
François 🙏 🙏 🙏

PS: This issue is all about AI, but I will mix things up going forward. 

The Marketing Intelligence Engineer: Your Future AI Secret Weapon?

With AI, deep and relevant context is essential. So much so that, at Augment Code, we made deep context understanding and our context engine the core differentiators of our AI coding assistant.

The better the AI understands a large codebase, its dependencies, and its context beyond the code itself (e.g., Jira tickets, Slack conversations, internal docs), the better the output of the coding AI. Without deep context, AI can produce high-quality code, but it will be generic. It won’t fit a software engineering team’s stacks, requirements, and styles.

That's why folks like David Nunez, former Head of Docs at Stripe and of Information Platform at Uber, foresee the rise of the Knowledge Engineer, a critical role on software engineering teams.

The same will be true in marketing: your marketing AI will only be as good as the context you feed it. Like Marketing enables Sales with full context, we need to enable our AI.

This responsibility may initially sit in Product Marketing, at the intersection of product knowledge, audience understanding, and messaging strategy. It will feed the AI with the proper positioning frameworks, messaging docs, brand style guides, competitive information, and more to enable the rest of the Marketing and Sales teams to derive value from AI.

Over time, this role may expand beyond context curation to also include:

• Selection and management of various AI platforms for marketing

• Information architecture for marketing knowledge

• Prompt engineering and curation for different marketing functions and tasks

• Expertise in connecting various marketing and data systems

We could then see this role evolve into full-on Marketing Intelligence Engineer and then move from Product Marketing to RevOps (or maybe IT in large groups?). It will partner closely with Product and Corporate Marketing to collect all that context.

It will become one of the highest-leverage roles in modern marketing departments.

How to get started:  

1. Begin by assigning context curation responsibilities to Product Marketing

2. Create a central repository of positioning frameworks, messaging docs, persona cards, battle cards, case studies, use case briefs, and brand guides for AI consumption (often the same as for sales enablement)

3. Assign someone in RevOps to partner with PMM to select, buy and set up your go-to AI tools

4. Link them to that context

5. Built GPTs, Gems or Claude Projects (see below) on top of them

6. If you don't know where to start, hire an AI consultant or trainer

My selection of 💡insights, 📰 news, 🛠️ tools, and ⚙️ workflows

⚙️ How I use Claude Projects as this newsletter’s editor

Speaking about giving AI the right context, here’s how I customized an AI editor that knows my newsletter's strategic brief, audience, and style guide.

This workflow should be applicable to most content you and your team create.

1. Created a Project in Claude (available in premium) to edit these highlights

2. Wrote a strategic brief (Google Doc). It includes “goals of the newsletter, target audience, guiding principles, preferred structure, reader feedback.”

3. Linked my Google Drive and Claude accounts

4. Added the doc as Project Knowledge (can’t attach Slides or Sheets yet)

5. Created a custom writing style by pasting 2 prior editions into a new style 

6. Created step-by-step instructions, so I don’t rewrite the same prompt every time": “Critique this draft. Propose a rewrite to adhere to the strategic brief and style. Include your rationales.” It makes Claude behave like a custom GPT.

7. Once I have a draft for an issue, I copy the draft in a new chat window in that project, select “extended thinking” mode and tell Claude to strictly follow my project instructions

🛠️ A cool use of Claude Artifacts

Lisa Adams built an interface that visualizes the potential evolution of each marketing function as we use more AI and agents. The point isn't that it's right, but it's pretty cool to see Claude’s output (use the drop-down).

Want to learn how to use Claude Artifacts? Here’s a good primer (free 60-min Codecademy course, beginner level).

💡Tackling the untapped: agents’ biggest opportunity?

Box’ CEO argues that agents will solve new use cases and enable work that simply wouldn't happen otherwise, instead of just automating existing workflows.

The competitive edge will come from imagining - and building - what was previously impossible. That's why understanding what AI is capable of is important.

🛠️ Speaking of giving context to AI: MCP, the universal connector for AI we need to become familiar with

MCP may become a technology as important as API for AI. It is already bringing essential context and tools to AI:

💡 How to prepare for a rapid transformation of work?

An AI trainer struck a chord when she offered her views on how to adapt if AI disrupts our jobs. She proposes (after we all first learn to use AI well):

My take: yes, these are important things to ponder. Things are going to change. How and how fast is the question. But Marketing skills will evolve rather than vanish. Let’s keep trying things out, leading with enthusiasm, learning with humility, and building remarkable relationships.

With appreciation. François | LinkedIn