Applied AI for Marketing Ops

Building secure, scalable, real-world AI workflows that actually work.

Hello! Welcome to my Substack, which I’m using as a blog to document how I’m building AI workflows to help teams work smarter.

I’ll be honest - I’ve been hesitant to start writing about my experiences and learnings, but after connecting with other AI builders (shout out to Justin Norris), I felt it was time to share. There’s so much AI hype out there and not enough practical, in-the-trenches implementation stories. This blog is my answer to that: the journey of making AI work with real business constraints.

How it all got started:

I’ve been in Marketing Ops for over 14 years, working at both enterprise companies and various startups, doing the usual Ops work: building MarTech infrastructure, optimizing processes, and building reports. For the last 3.5 years, I’ve been leading a team that manages our MarTech stack and global campaign execution.

When ChatGPT exploded onto the scene, I was using it like everyone else - to help write better emails, brainstorm campaign ideas, etc. But then our leadership challenged us: “How can we use AI to make our sellers more efficient at account research?”

My Ops brain went into overdrive. All those years building workflows, using tools like Zapier and Workato culminated in a single obsession: What if I could automate all of that research?

So I started experimenting. I became hyper-focused on using AI to automate hours, even days of mundane research. Things like pulling account 10Ks, annual reports, org charts, and earnings calls, then extracting key insights and mapping them to our case studies - all with one click.

And on top of that challenge? Building this entire tool to be secure and compliant within our Azure environment. I even added automated PPT generation and email follow-up sequences. (I’ll do a breakdown of that project in a future post).

That project opened the floodgates of even more tools I started building: AI-created ABM assets at scale, automated outreach sequences, Asana workflow automation, and campaign reporting via AI. The possibilities were endless.

What I’ll be sharing:

I’ll be sharing my journey and various learnings from building AI automation that works, with real constraints: enterprise-grade security requirements, Azure and Microsoft infrastructure and compliance frameworks, which makes seemingly easy AI “hacks” or simply dumping everything into ChatGPT impossible. (Although I use all the major LLMs in different ways, so I won’t be entirely focused on Microsoft environments.)

Here’s what you can expect:

• Applied AI workflows that adhere to enterprise security, within Azure and Microsoft environments.

• What doesn’t work - my next post, for example, is on how multiagent workflows can be over-engineered traps compared to simpler solutions. And the lessons from AI workflows & tools I’ve built and abandoned.

• Real data & processes from AI-driven campaigns and automation.

I welcome any thoughts, ideas, or questions - this AI world is moving faster than any of us can keep up alone. Let’s build and learn what actually works together.