Jeff Canada:
All right, here we go. Hi everybody. It’s really great to be here. I was telling some people earlier that I’ve been coming to this event since it started at True Normand. It was right around the time that I moved to San Francisco, so it’s really cool to be up here and to be talking to you today.

Let’s get into it, because we don’t have a whole lot of time. Real quick agenda. I’m going to introduce myself, talk about demystifying AI for MOPs, get into a couple of use cases, and then a few lessons learned to wrap it up.

So first, my favorite topic, which is me. It’s not really. I have been in the marketing space for longer than I care to admit. I’ve worked across the world, in London and New York, and most recently here in San Francisco, and in a bunch of different industries—finance, clinical trials, technology, and pricing for petrochemicals. I really started to focus more on the operations side of things over the last 10 years since I’ve been in San Francisco.

My current role is that I lead our B2B marketing operations efforts. My title changes every day, but this is generally what I do. I look after our B2B MOPs. And outside of robots and technology, I love to bake bread. I love my dog Scout—he’s going to make an appearance, maybe not here since he’s sick, but on the slides. And like Kim said, I am a sound healer, so I absolutely love music, making music, and traveling to see music. I was just in Pasadena this last weekend for a concert.

But enough about me. Let’s talk about what you guys are all here to talk about, which is AI. It’s really hard today to open your phone, get on your laptop, or go to a conference without hearing about AI. AI is transforming everything. It’s changing the way social media works. It’s changing the way journalism works. It’s changing the way marketing works. It’s everywhere.

But what we’re starting to see now are reports that AI project failures are on the rise. I think I read something from MIT recently that said 97% of AI projects are failing. So is it taking over everything? Is everything failing? It’s confusing.

And when we get to the reality of how people are actually using AI—it’s Scout. It’s making pictures of our dog dressed up like Pikachu. The big viral moment we had after ChatGPT launched was image generation and people using it for fun. So yes, there’s potential for it to change everything, but right now, we get Scout as Pikachu. Poor guy has kennel cough.

What is possible? A lot. I’m not going to run through the full list, but there’s quite a bit from a business and marketing ops perspective that we can do with AI—scientific exploration, code review, data mining, and more. But we’re stuck generating dog pictures, right? So how do we get from doggy Pikachu to AI transforming business?

I was at a dinner a couple of weeks ago sitting next to a woman named Kristin from Scale Ventures. She was the AI Executive in Residence and was talking about levels of AI maturity. Who loves a maturity model? I sure do. She was talking about three levels of AI adoption, and I thought, that’s brilliant. I said I’d only say the F-word twice, and I’ve already said it three times, but it was brilliant. I wanted to adjust my presentation to talk about it that way because, like a game board, you can’t get from start to finish without going through the levels.

Personally, I’ve been using AI to get from just typing stuff into ChatGPT to actually building real AI automation. I think of eight core competencies that AI helps me with in my day-to-day: content generation, help desks, research, coding, data analysis and management, ideation, translation, and task automation. I’ll walk through use cases that touch on some of these.

The first level in AI adoption is LLM-driven. That’s where I spent the first year—using ChatGPT separately from my actual work. It was frustrating but helped me learn what works and what doesn’t. Each competency helps me be better at my job.

Content generation—one of my favorites—is great for summarizing meetings or translating technical content for a CMO. Anyone else struggle with that? It’s painful sometimes, but ChatGPT helps.

Help desks—when I evaluated a new marketing automation platform, I used AI to handle repetitive vendor questions. I was a team of one, and it saved time. I have a team now: two humans and one ChatGPT.

Research—I use it constantly. Coding—writing SQL queries or using App Scripts for automation is now possible without engineers.

Data analysis—I ask AI to check what I’m seeing in the data and confirm or challenge it. Ideation—it helps overcome blank-page syndrome. Translation—it prevents embarrassing mistakes in global communications. Task automation—my colleague Steve built a UTM builder using a custom GPT. It replaced a messy spreadsheet and saves hours.

Level one is using LLMs like ChatGPT. Level two is using AI within tools you already have. You don’t always need new tech—ask vendors like LeanData how they’re integrating AI. I used Clay for data enrichment at OpenAI and layered in AI to analyze form submissions. For example, a Spanish-language submission that our system might have rejected was actually a great enterprise lead once AI summarized it. That small change led to hundreds of millions in revenue.

The third level is full AI automation. Before you dive in, ask yourself, “Do I need AI or just automation?” Sometimes cleaner data is the real answer. My AI assistant, which I call Mops-O-Matic, was born from that.

I had a colleague send me a lead upload request at 11 p.m. on a Saturday. It took me almost an hour to clean, format, and import. So I automated it. I built Mops-O-Matic to take a spreadsheet, ask questions, and format and upload everything into Marketo automatically. The process went from 57 minutes to 57 seconds.

The lesson? Start small. I tried to do too much at once and failed at first. If your data or process is bad, AI will just multiply the mess. It’s a multiplier. Clean data and strong processes matter more than ever.

Start with one use case. Iterate. Build incrementally. AI isn’t here to take your job, it’s here to elevate it. I don’t want to spend two hours renaming and formatting lists—I want to build smarter systems. I had to build automation to get time back to build more automation. Now I can focus on more valuable, creative work that drives business impact.

Speaker 1:
I think that’s it. Thank you.

Jeff Canada:
Thank you. That’s my email. Send me an email or find me on LinkedIn. If you message me there, say that you were here—I have thousands of requests. If you try to sell me something, I’ll scream and unfollow you.