From Hackathons to an AI Content Engine: Sylvia LePoidevin on Iru’s Marketing Transformation

This week on AI’ve Got Questions, Stacey sits down with Sylvia LePoidevin, CMO of Iru (formerly Kandji), to talk about how she’s turning AI from a buzzword into a company-wide movement.

Sylvia shares how she built an “AI-first” marketing team, starting with a scrappy hackathon that produced a working content system in just four hours, to a company-wide “Marketing Shark Tank” where bold, AI-powered ideas compete for cash prizes.

They dig into:

• How to make AI adoption fun instead of intimidating
• The rise of “tastemakers” and “operators” on modern marketing teams
• Why every brand should build a human “anchor” for its content flywheel
• The surprising SEO “hack” that boosted Iru's AI discoverability

If you’ve been wondering how to bring your team along on the AI journey, and actually make it exciting, this episode is for you.

Transcript

Stacey Epstein:
Before we get started, a quick note — the company you’ll hear us talking about today recently changed its name. It was known as Kandji, and it’s now called Iru. Same great team, new name, and an even bigger vision. So when you hear us mention Kandji, we’re talking about Iru. Let's go!

Welcome to the show. I’m thrilled to have Sylvia LePoidevin, the CMO of Kandji, join us. We met several years ago when you were early in your Kandji journey. I heard about Kandji, became an early investor, and I’m proud of that. You’ve done so many amazing things over the years, and now there’s a lot of cool work with AI. I can’t wait to get into it. But first, tell us about your journey. How did you get to where you are now?

Sylvia LePoidevin:
Thank you, Stacey. I’m excited to be here. My path is pretty unconventional, which I think shapes who I am and what I’m building. I grew up in remote parts of Africa and moved to the U.S. at 17 for college. I fell into tech early. At 22 I joined a startup as employee number 12; they were literally working out of a house. I loved building something from nothing—the energy and pace of tech. That’s only accelerated with AI, of course. My career has been a series of zero-to-one stories as the first marketer, building from scratch.

That was my journey at Kandji, too. I joined a little over six years ago as employee number four. We’re now nearly 300 employees with an ~$850M valuation. It’s been an incredible ride with a lot of exciting AI projects ahead.

Stacey:
You’ve been a very important part of Kandji’s success. Most of our listeners are marketers and know the value of great marketing, but when it’s done right, it can truly fuel a company’s growth. Give us the quick Kandji overview, then we’ll come back to AI.

Sylvia:
Kandji manages and secures Apple devices for companies—think MacBooks for employees, iPads in retail, and more. We typically sell to IT and security teams at organizations using Apple. Tech companies are a big segment, but use cases go well beyond that. Apple at work is only becoming more common.

Stacey:
Same. I don’t use anything but Apple. In my line of work I still get Word docs and PowerPoints and I’m like, people, can we get more modern? Don’t attach files and send them to me. I’m a big fan—and, as I said, an investor—so I’m excited to see the growth.

Let’s talk AI. GenAI hit a couple years ago and surprised everyone. Season one of my podcast focused on AI tools for marketing, but CMOs kept asking me, “How do I actually run my team with AI? How do I incorporate it into everything we do?” There’s pressure to adopt AI without a roadmap. What did you do when ChatGPT landed? What were your first steps?

Sylvia:
The first big move—about eight months ago—was an AI hackathon. It was intentionally unstructured. We booked four hours in a conference room and set a goal: build something that automates content distribution and repurposing. That’s a classic low-hanging-fruit use case for AI, though it can be done well or poorly.

We split into three groups:

1. Knowledge base—how to give AI the context it needs to write accurately.
2. Format and taste—what “good” looks like for us: tone, voice, structure, word choices.
3. Operations—connecting systems and designing the ongoing process.

We built a custom GPT and, by the end of the session, had a working system. I’ve heard people say prompt engineering should be called “context engineering.” Without the right context and taste, you get average-of-the-internet output. If you just churn out more average content, it won’t matter.

Stacey:
Did you know ahead of time that the hackathon would focus on content?

Sylvia:
Yes. We chose it going in because it was high-impact and achievable in a single session. The goal was to walk out with something usable—and we did.

Stacey:
So, a content machine. Did you change team roles after that? Do you now have dedicated people for those three areas?

Sylvia:
Not formal titles, but we’ve seen people naturally fall into two camps. Tastemakers—they shape the knowledge base and “what good looks like,” often with storytelling, film, or journalism backgrounds. And operators—they love building systems and automation. Pairing a tastemaker with an operator produces great results. Some folks do both, but most lean one way. We didn’t change titles, but roles and responsibilities evolved.

Stacey:
What results have you seen? More content? Better content?

Sylvia:
Both, but the important shift is launching new human-led content streams while using AI for repurposing and distribution. I recently spoke with a CMO who said, “We went from three blog posts a month to five,” and I thought, why? Who cares? We use an anchors and distribution model.

The anchor is a human, high-quality content source—for example, our podcast “Patch Me If You Can.” It’s a real conversation that becomes the engine for the flywheel: social clips, derivative articles, topic extraction for new pieces, and more. The distribution is where AI shines. In my view the anchor should not be AI-generated; that spark comes from lived experience.

Stacey:
I love the anchor idea. The original content can be deeply human, then AI helps you multipurpose it for different audiences and formats. I don’t even do enough of that with my own show. It would be easy to spin out more.

Sylvia:
Totally. We also launched a media site called The Sequence because we needed a home that felt like its own publication rather than “the Kandji blog.” It houses our podcast conversations, reported articles based on interviews, original research from audience surveys, and short product “wow moment” videos. AI-powered repurposing lets us support multiple streams while keeping the source human.

Interestingly, The Sequence lives on a separate domain. A month or two after launch, we noticed LLMs were citing it a lot. Our theory: when people ask AI tools questions, those tools look for answers on our site, then try to validate with external sources. Because The Sequence sits on its own domain, it can be treated like a third-party reference. That wasn’t the reason we created it, but it’s become a strong discovery channel. Early data shows a higher-than-expected number of visitors finding us through LLMs.

Stacey:
That separate-URL “hack” matters right now. Authority is everything. What other AI-driven ideas are helping inbound and pipeline?

Sylvia:
We ran a Marketing Shark Tank Day—an internal event for bold ideas where every pitch had to include a transformational use of AI. Our “sharks” were internal folks role-playing our buyer personas—IT and security leaders—so we could pressure-test ideas against what actually lands with buyers. With AI hype, a lot of teams forget the customer and just increase output. We wanted the opposite: use AI to improve the buyer experience.

It’s early, but standout ideas included a digital twin of our solutions engineering team so prospects can tap that expertise on demand, plus interactive experiences that add delight and clarity. The goal is to meet buyers where they are, not do AI for AI’s sake.

Stacey:
I have two ideas for that SE digital twin. Past guests: 1Mind (Amanda Callow) for a live AI persona over your knowledge base, and Ovation for video navigation where users converse with your demo library. Also, I love that both examples—the hackathon and Shark Tank—were internal. They’re great for culture and recruiting, and they take pressure off the CMO to own every AI idea. Invite the team in, flatten the org, and pick the best ideas together. That involvement is energizing. I bet Shark Tank Day was a blast.

Sylvia:
It was. We had cash prizes, music, and way too many hilarious AI-generated visuals. So much fun.

Stacey:
I love it. In past roles we manufactured moments like that too. What’s exciting now is that AI gives everyone a way to contribute to real transformation, not just campaign ideas. Bringing in frontline voices surfaces ideas leaders might never think of.

Sylvia:
Exactly. Top-down AI mandates can create fear. If I could wave a magic wand, I’d remove the fear so people feel free to think big. When we create space to learn from each other, the collective knowledge is surprising. No one is formally trained in this; we’re all figuring it out. Give people room, reduce the fear, and amazing ideas surface—especially from those closest to the work.

Stacey:
Amazing. I loved catching up. Let’s check in again soon. And for anyone listening, if your company uses Apple devices, send your IT team to Kandji or to the content site, The Sequence.

Thanks for coming on the show.

Sylvia:
Great to be here. Thanks, Stacey.