AI Isn’t Killing Our Human Creativity. It’s Empowering It.
Like a lot of agencies right now, we've been figuring out where AI actually fits into the work. It goes without saying that these new tools are revolutionary. That said, it’s easy to confuse hype for practicality and actual utility in a process built on craft, instinct, and human collaboration.
At Herd of Shepherds, we've landed on two specific tasks we’ve found ourselves leaning on AI to do more of, and they've quietly reshaped how we do what we do.
The first is de-risking bold ideas by using generative tools to make unconventional concepts feel real before a client has to take a leap of faith on them.
The second is building technical systems that would've been out of reach for a shop our size a few years ago by using AI to do the kind of heavy infrastructure work that used to require a much bigger team.
Neither of these is about handing the work off. Both are about clearing runway so the human part (ideas, the craft, the collaboration) can really take off. Here's what that looks like in practice.
1. De-Risking the Weird Ideas
Swansons wanted to make gardening feel fun and approachable for beginners. So, we proposed a series of quirky folk songs about dirt, herbs, and colorful plants, performed and captured in a single dolly-shot inside the nursery.
On paper, it sounds un-pitchable. The kind of thing that, described in words, risks sounding like a Portlandia sketch.
So we didn't just describe it. We used Midjourney to nail the visual tone, ChatGPT to experiment with lyrical directions, and Suno to generate a placeholder banjo track that brought the whole thing to life. None of it was final. But it was close enough to move the idea from "out there" to "oh, we get it."
There’s a much-researched psychological reason why that matters. When people encounter something genuinely new, their brain hits a wall of cognitive dissonance. The idea doesn't fit any existing mental model of what a "good ad" looks like, so the safer move is to back away. But give the brain something concrete to hold onto — a rough track, a mocked-up scene — and the dissonance drops. The idea stops being abstract and starts to be evaluated as a legitimate option.
That layer of AI unlocked the opportunity to go make the real thing. We went into music production on Bainbridge Island with producer Johnny Bregar We hired actor singer/musician Chloe Hendrickson from Nashville and bring in director Tony Fulgham to help bring it all to life in font of the camera. The result was something no algorithm could have manufactured. It was warm, weird and very much human.
AI didn't make the campaign. It made the campaign feel “buyable.”
2. Building Efficient Systems That Scale
Smartsheet came to us with a unique challenge. They wanted an enterprise campaign that went deep on technical features, real workflows, and role-specific value. They wanted content that actually spoke to the people using the platform within specific use cases.
We needed to build a system complex enough to serve all those roles and use cases, but structured enough that a brand editorial team could pick it up and run? We needed to create and authoritative expert they could lean on.
Our solution was a Retrieval-Augmented Generation (RAG) system using the OpenAI API, fed with a mountain’s worth of Smartsheet's enterprise documentation, all wrapped in a simple, user-friendly, prompt interface. In other words, a private AI assistant trained on Smartsheet's own knowledge base, ready to help writers and content creators produce technically accurate work without becoming SMEs themselves.
Upon deployment, content development time dropped by 5x and the pages created outperformed every previous Smartsheet enterprise paid campaign. And the client-agency collaboration hit a new level, because we were all working from the same source of truth.
The AI system we built didn’t start creating the campaigns. It just cleared the way so humans could work better and faster than ever before.
So, where does that leave us?
There's a lot of justified concern about AI replacing jobs, reducing craft, or flooding the world with noise. Our experience has been the opposite. AI hasn't replaced the work. It's helped us sell bolder ideas and build smarter systems, so we can spend more time doing the part that actually matters: making real things with real people.
AI hasn't stolen our magic wands. It's given them new powers.