If you are still delivering branding work in a PDF, you are doing it wrong.  

How the team turned a universal creative handoff problem into a living, breathing brand messaging tool for consistent, audience-aligned copy.

There's a scenario every brand team knows too well: a messaging strategy arrives polished and full of promise, only to slowly mutate as it passes from agency to comms, marketing, and external vendor, until it barely resembles itself. For Kettle's strategy, creative, and product teams, that problem was too familiar to ignore. So they built something to fix it.

We sat down with Brian Austin, our Group Director of Strategy, Britta Jerome, Senior Strategist, and Jonathan Hooker, our Director of Technology, to talk about what they made together, and why the best ideas sometimes start with a simple "what if?"

Let’s start at the beginning, what was the original problem you needed to solve?

Brian: There's a really universal problem at the core of it: how do you take something that feels intangible, like a messaging strategy, and make it actually functional and usable day-to-day? The work we do as strategists can feel a bit hypothetical. A creative team receiving a polished deck still has to translate it into real artifacts, copy, headlines, campaign language. That translation is where things start to unravel.

Britta: It can feel like a really long game of telephone. By the end of it, the work is saying something completely divorced from the original strategy. Every branding brief we get seems to start with, "We did this exercise two years ago, and somehow we've undone everything." That's a very real scenario we see constantly. This tool was about stopping that clock.

So how did the idea for an interactive, AI-powered bot specifically come about?

Jonathan: We didn't start by saying, "Let's build a bot." We started with the problem. In our first working session, we explored a few directions, a more intuitive website, Figma plugins, and different interactive experiences. But pretty quickly it became clear that a bot-like experience with a simple web interface, one that lets you toggle between personas, channels, and use cases, felt the most additive.

Britta: While working through the messaging framework, we almost couldn't help ourselves. "What if we had something to complement this? I wish I could just interface directly with this document." And then it was like... well, what if you actually could? It felt like such a natural next step.

What did translating a strategy document into a working tool actually require?

Jonathan: The guidelines were already written in a very "do this, don't do that" format, which lent itself beautifully to building a rule set the bot could check against. The primary technical task was structuring that data, getting it into a spreadsheet, quantifying certain things so the system could handle multiple options, and referencing back to the original document so that when it flagged something, it could point you directly to the relevant page and the relevant persona.

Britta: On the strategy and copywriting side, it was a fascinating exercise. You've written this full document where everything lives cohesively, and then you have to break it apart into individual pieces. Does this rule actually work in isolation? Does it conflict with another rule? How do you handle an audience persona that rubs up against a different one? Taking something that big, breaking it back into pieces, and then watching it come together in a new way was genuinely fun.

How did you make sure the tool felt useful rather than robotic?

Jonathan: We started with what felt right from a technical standpoint, but you can't really validate that in isolation. So we brought copywriters in to run real copy through it. The early feedback was that it could be overly critical, or overly positive. So we started tuning the rules to allow for more nuance: not just yes or no, but "That's close, here's what you might want to consider to align with these other guidelines." There's rarely a sentence that's perfectly on-brand in every dimension at once.

Brian: That's something we talk a lot about in the context of AI tools, keeping the human in the loop. The way we built this, with copywriters testing, giving feedback, us improving it, and cycling through that a few times, that's what actually made it useful for real teams. It's a testament to where we are with these tools: the best of technology and the best of human expertise working together.

How did we ensure this was truly complementary to the strategy work, not separate from it?

Britta: Transparency. One of the things we always appreciate as AI users is when a tool doesn't just say "this is wrong,” it shows you where and why. The bot doesn't just give a thumbs down; it points back to the document, to the specific rule, to the persona. That connection back to the original thinking was really important to us. We didn't want it to feel like a black box. We wanted it to feel like a knowledgeable collaborator who had read the same brief you did.

Jonathan: And that's really where the product and the strategy converge. The messaging work had to be airtight before the bot could be useful. They're interdependent. You can't build a smart tool around a vague ruleset, and a brilliant strategy means nothing if people can't access it when they need it.

What can both brand and agency teams learn from what you've built?

Britta: That the solution to a strategy problem doesn't always have to be another document. There's this classic idea that disappointment is the distance between expectation and reality, but it can go the other way too. When a client is genuinely surprised by how useful something is, that's the payoff for everything that got you there: the problem-finding, the feedback loops, the iteration. That delight is the whole point.

Jonathan: For me, it's that strategy and product aren't separate tracks. When they're in conversation from the start, when a rule in a messaging framework has to survive being built into something functional, everyone's thinking gets sharper. The work gets better. That's what this project showed us.

If your team is struggling with similar messaging challenges, or this type of tool sounds like it could be useful, get in touch with us!