Mitch Hislop | VP, Group Digital Director
One year ago, we launched our proprietary AI tool, Butter, with a simple goal: to make our work smoother, faster and more creative. What started as a playful experiment quickly became part of our daily workflow. We’ve built custom GPTs, ran thousands of sessions and learned more than we expected about what AI can (and can’t) do when paired with human brains. Here’s what we discovered.
Why we built Butter
When we decided to invest in AI and build Butter, our goal was to empower our teams to focus on what matters most: delivering extraordinary work for our clients.
Recognizing the increasing pressure on marketers to be faster, more efficient and more innovative, we embraced generative AI not as a trend, but as a transformative tool. With Butter, we can streamline time-consuming tasks, freeing our people to spend more time creating, strategizing and engaging. It’s not just about productivity — it’s about making space for creativity and meaningful impact. Going all-in on AI was an easy choice.
What we learned (sometimes the hard way) building custom GPTs
Early on, we noticed we were typing the same prompts over and over again. We needed consistency. We wanted speed. We wanted tone and voice that reflected our brand and our clients’ brand, not boilerplate content. So, we started building custom GPTs tailored to our work.
What began as a few quick-fix solutions turned into a shared resource library that powered everything from creative briefs to brand naming exercises.
A few key learnings:
- Prompts are process. The better we understood our own workflows, the better Butter performed. We turned vague ideas into structured prompts, which turned into repeatable systems.
- Collaboration > perfection. Butter rarely gives you a perfect result on the first try. But it makes an incredible partner in early thinking and ideation.
- Build for your bottlenecks. We focused on areas where we were getting stuck: compiling metrics, writing briefs, summarizing strategy, breaking through blank-page paralysis.
Tangible process improvements and our favorite GPTs (so far)
Benchbot: With the ability to write SQL, talk to our Benchmarks database, and analyze data, Benchbot lets anyone at broadhead turn data into insights, without needing to tie up someone on the Data Science & Analytics team.
Image Resizer: Resizes uploaded images to user-specified size and format.
Dynamic Persona GPTs: Allows our team to chat “with” our personas, test creative ideas and easily access knowledge about the persona or client.
Butter added to our culture, too.
Using AI wasn’t just efficient—it was fun. The team got weird with it, and we loved that. Here are just a few of the ways our team is using it outside of work:
“I built a custom GPT to manage my entire knitting inventory. It knows every pattern I own, the tools in my kit, and the type and quantity of yarn I’ve stashed away. Now, I can find projects I already have the materials for or check in real-time whether I have the right needles while I’m standing in a yarn shop and can’t remember what I own!” — Maren Viker, Senior Strategist
“I gave Butter the random assortments of ingredients I had in my kitchen when I couldn’t think of anything to make out of it, and Butter gave me a recipe to use up some of the ingredients. Worth it.” — Julia Miller, Agency Marketing Intern
“I adopted my dog at three years old and never got puppy pics, but after dropping a few current pics in, I asked for a good guess for what she looked like as a puppy and it gave me the cutest animations.” — Leah DeMars, Senior Account Manager
What’s next?
A year in, Butter is less of a tool and more of a teammate. We’re planning more custom builds, smarter integrations and a few new rituals to keep us exploring. We’ll keep asking weird questions. We’ll keep making the work better.
Because we’re not just working in Butter — we’re building with it.
PS: Check out our Butter (Un)wrapped infographic to see what a year of AI-powered collaboration looks like:
