Vibe Coding
for Marketers

Vibe Coding for Marketers: How to Stop Prompting and Start Building

Most marketers treat AI like a chatbot. Ask a question, get an answer, copy-paste it somewhere. But the teams pulling real leverage aren't prompting. They're building: landing pages, analysis scripts, research systems, automated workflows. No code required.

In this Grow & Tell session, Sarah Charlton (AI GTM Strategist, former Growth & Marketing leader at Asana and Dropbox), Austin Hughes (co-founder and CEO at Unify), and Quintin Smith (Product Marketing at Unify) showed how marketers can move from casual AI usage to building real tools. Live demos, live failures, and a 90s roller-rink theme included.

AI Is Your Analyst, Not Your Strategist

Sarah and Austin opened with a framework for where AI actually delivers for marketers: quantitative analysis that would have required a dedicated analyst, synthesizing unstructured data like sales call transcripts at scale, spinning up fast MVPs for campaign experiments, and generating volume when you need 100 ideas to find the 10 worth pursuing.

Where it consistently falls short: strategy, taste, and copy. AI can't tell you what's worth doing. It can't write thought leadership that doesn't read like AI. And the 12-word headline that compels action? It doesn't understand the psychology.

Sarah put it bluntly: "Claude will say, 'you're done, just ship this now.' And I'm like, what are you talking about? This is not done."

Austin added the nuance on voice: "Never once have we just taken a post that came out of AI. It's not interesting. It doesn't understand what's happening in the zeitgeist, how people are feeling about things. That's gonna take a lot to get there still."

Deep Research Without the Tab Explosion

Sarah introduced Genspark, a multi-agent research tool she called "criminally underknown in the United States." Unlike single-model chatbots, Genspark orchestrates multiple AI models and specialized tools like video transcription to tackle complex research questions.

Three live demos showed what this looks like: comparing Sam Altman's and Dario Amodei's AI visions by actually transcribing and synthesizing their recent podcast appearances (not just summarizing summaries), running narrative research to prep Austin for an upcoming podcast by mapping the stories shaping the sales-plus-AI conversation, and analyzing competitor homepages, blogs, webinars, and keynotes in parallel for a client in the AI ERP space.

Genspark's agents check each other's work, which cuts hallucination. The tradeoff is weaker writing. Sarah's workflow: research in Genspark, then feed findings into Claude for the actual prose.

From Wireframe to Working Landing Page in 15 Minutes

The centerpiece: Sarah live-built a Unify ROI calculator landing page using Lovable, voice dictation via WhisperFlow, and a wireframe Quintin sketched for her in under 30 seconds.

Her vibe-coding framework: start with a vision (don't brainstorm inside the coding session), define how the logic works (feed it the spec, don't ask AI to guess), build in chunks (hero, calculator, quotes, CTA, each with its own prompt cycle), and talk instead of type. "When you talk, you don't censor yourself as much. You're more expansive, you give more context, and you move faster."

After 15 minutes: a branded page with Unify's actual logo, a working calculator with input sliders, real customer quotes with photos pulled from Unify's site, and a toggle switching everything into 90s roller-rink "fun mode."

Sarah was honest about where it landed: "First reaction is gonna be, this page sucks. And that is a fine starting point." The point isn't first-pass perfection. It's the compression.

Quintin framed the shift: "I've gone through this process before, building an ROI calculator. Getting the data, framing the calculations, working with a developer, working with a designer. That's usually months of work. Now it's hours with a single resource."

Turn Your Ad Spend Into a Python-Powered Audit

Sarah shared her framework for ambitious, multi-step marketing projects, using SEM optimization as the example.

Three phases: research with Claude Chat (understand best practices for your specific situation, then have it generate the prompt for the next step), build with Claude Code (feed that prompt in and let it write Python scripts that analyze your actual campaign data and produce recommendations), execute with humans (scripts produce recommendations, not actions; people decide what's worth implementing).

Sarah's warning on the handoff: "Don't just dump this into Claude Code. Did it pull from weird listicles instead of actual expertise? Use your judgment on what comes back."

The research-then-build-then-execute loop works for any complex analysis project, not just SEM.

How Unify Turns Brainstorms Into Automated Workflows

Austin walked through how Unify converted a brainstorming session with Sarah into systems that run every week. A weekly agent processes every sales call recording and tracks competitor mention trends over time. Webinar operations that used to take 10+ hours per event now run in the background. An agent mines Slack for LinkedIn post ideas, though the posts themselves are always written by people.

Austin's framing: "The kernels of insight need to come from people still. AI gets you from those first 2 or 3 ideas to the 7 through 10. And instead of relying on our own memory bank of everything happening, we can just do that programmatically and spend more time on the work that matters."

Start Small, Stay Curious, Check Back Often

Austin's closing advice: "The intuition is built from trying things, not from reading about them as a third party. Start with small things. It doesn't need to be perfect."

Sarah added that play makes hard things easier to learn: "If you're really serious and you've got to one-shot it in the next hour, you don't have the mental space to explore." And one last tip: tools improve faster than you expect. The product you dismissed three months ago might now be exactly what you need.

About the Speakers

Sarah Charlton is an AI GTM Strategist and former Growth & Marketing leader at Asana and Dropbox. She writes about AI's impact on go-to-market on her Substack and LinkedIn, and builds AI tools and skills for marketing teams.

Austin Hughes is the co-founder and CEO of Unify, where he's building the platform for signal-driven outbound.