Maze

Making developers trust an AI agent's judgement enough to act on its fixes.

SaaS
B2B
Launch
The Problem

Maze were launching the remediation side of their AI agent. Not just the part that flags a vulnerability, but the part that reasons through it and tells you how to fix it. That's a much bigger ask than a normal dashboard feature. You're asking developers to trust the agent's judgement enough to actually act on what it says, and that's a hard sell to an audience who are naturally skeptical of "AI does it for you" claims.

So the launch had two jobs. Get attention fast on social. And hold up under the kind of scrutiny developers give anything AI related, because their first question is always going to be "how does it actually reason through this, and what data is it looking at to get there."

The Approach

Instead of listing features, I built the whole video around the agent itself.

You follow it through one journey from problem to fix, so you're watching it think rather than just being told it's smart. That was the call I made early on: trust comes from showing the reasoning, not from claiming the capability.

I kept the top level messaging simple, but packed real detail into the visuals underneath, the actual signals the agent reads and the boundaries it works within. Developers don't take "AI powered" at face value. Give them something they can actually pick apart and they'll believe you a lot faster.

I kept the visual style close to Maze's existing branding and pushed it into motion for the first time. Modern and technical enough to hold credibility with developers, but with a more fluid, organic movement for the agent itself, so it felt like something with judgement rather than a script running through steps.

The other big call was distribution. This was Maze's first animated brand asset and social was always going to be the main channel, which meant fast moving and easy to scroll past. So I never treated this as "one video." It was a teaser to build anticipation before launch, the hero video to carry the story, then a full set of cuts and loops so they had material to keep using long after launch day, across whatever platform they needed to show up on.

"The animation both looked good and it gave our audience a way to immediately understand what we were launching. And, we got way more than just a single video, meaning we can show up everywhere from day one."

Maze
Pre-production

Scripting the trust angle. Working out what to show and prove, and in what order, before any animation started.

Animated Teaser Video

A short cut released before launch to build anticipation and get people asking questions ahead of the full story.

Social Cuts

The launch story re-cut for how people actually watch on social. Fast, sound off, thumb stopping, without losing the logic underneath.

Distribution Pack

Thumbnails and loops built to work across email and embeds, not just the main post, so they could drive plays from every channel.

Results

A launch that did more than announce a feature. It gave Maze a way to explain their AI's reasoning to a skeptical audience, and content they're still using well past launch day.

Want to take your launch to the next level?

Learn more about the launch package