Vibe Coding as a Marketer: 6 Tools I Built with AI

I do not have a computer science degree. I have never taken a formal programming course. Until about a year ago, the most technical thing I did at work was write Excel formulas and wrestle with HubSpot workflows.

Now I build internal tools.

Not because I suddenly became a developer — I did not. But because AI coding assistants, primarily Claude Code, made it possible for someone with my background to describe a problem in plain language and end up with working software on the other side.

This is what people mean when they say “vibe coding.” You are not writing every line yourself. You are thinking through what you need, communicating it clearly, reviewing what gets produced, and iterating. The judgment is still yours. The typing is shared.

Here is what I have actually built, and why it mattered.

Why I Started#

The trigger was frustration. Our B2B sales team was using a generic CRM that had no concept of how our industry worked. Fields were wrong. Workflows made no sense for how we actually sold. IT had a backlog, so a custom build was not happening anytime soon.

I had been experimenting with Claude Code for writing tasks. On a slow afternoon I described the problem to it, explained what the sales team actually needed, and started asking it to build something. Three days of evenings later, we had a working internal CRM with custom fields, display logic built around our specific sales motion, and an interface the team could actually use.

That changed my understanding of what was possible for me, specifically.

The Tools#

1. Custom B2B CRM

Built for our sales team based on how our industry actually operates, not how generic SaaS assumes it does. The system includes industry-specific fields, a display structure that matches how reps review accounts, and enough flexibility to adapt as our process changes. It is not replacing Salesforce. It filled a gap that Salesforce was not filling for us at our scale and budget.

2. Red Dot Design Award Scraper

We track design trends in our product category. The Red Dot Award is a useful signal for what directions premium manufacturers are moving. I built a scraper that pulls winning products by keyword, so I can monitor shifts in materials, form factors, and positioning without manually combing through the site every quarter. This feeds directly into how I brief our own product messaging.

3. Industry Association Lead Scraper

Member directories from industry associations are some of the most underused lead sources in B2B. The contacts are self-identified, the company profiles are often rich, and the intent signal is implicit. I built a scraper that collects member names, contact information, and company details from these directories. What used to take an intern a week of copy-pasting now runs in an hour.

4. Trade Show Exhibitor Scraper

Same logic, different source. Trade show exhibitor lists tell you who is actively investing in a category, which is a meaningful buying signal. The scraper pulls exhibitor information and listed contact persons, which goes directly into our outreach pipeline.

5. Print Cost Estimation Tool

This one solved a specific operational headache. Our team frequently needs rough printing cost estimates early in a project, before we have full specs. I built a tool where you upload an image of a proposed piece, adjust parameters like size, quantity, material, and finish, and get a working estimate back. It is not a substitute for a real vendor quote, but it is fast enough to make early-stage decisions without going back and forth with the production team every time.

6. Zi Wei Dou Shu Chart Frontend

This one is a side project with a different audience in mind. Zi Wei Dou Shu is a traditional Chinese astrological system with fairly complex calculation logic. I built a frontend that renders the chart and, usefully, exports the underlying JavaScript parameters in a customizable format. The export feature is what makes it practically interesting — it allows the output to connect with other tools or workflows rather than just displaying a static result.

What This Workflow Actually Taught Me#

The skill is not coding. The skill is scoping.

Every tool I built started with me being very specific about what problem I was actually solving and what the minimum usable version looked like. Vague prompts produce vague tools. The cleaner my brief, the faster things moved and the fewer dead ends I hit.

I also learned to manage the process, not just the output. I use a tool called cc-switch to manage API token budgets across different tasks, which matters when you are running multiple projects. I use structured agent workflows — what I think of as guided decision-making frameworks — to break complex tasks into steps that are easier to review and course-correct. The AI handles generation. I handle direction and judgment.

The third thing I learned is that documentation is a legitimate output. Several of these tools required me to work through dense technical documentation to understand what was possible. I started writing that down in plain language, turning it into guides I could reference and share. That process of translating complexity into accessibility is, it turns out, a marketing skill that transfers directly.

What This Means for Marketing Going Forward#

The gap between “someone who understands the business problem” and “someone who can build a solution to it” is narrowing fast. That gap used to require hiring a developer, filing a ticket, or waiting for budget. Now it sometimes requires a Saturday afternoon.

This does not make developers unnecessary. It makes business-side people more capable of acting on what they know. A marketer who understands lead generation and can build a scraper is more useful than one who can only brief a developer on what they want. A growth person who can prototype a cost tool is closer to the problem and faster to iterate.

I find this genuinely exciting — not because AI is impressive as a technology, but because it changes what someone with my background can contribute. I have spent my career moving between psychology content, B2B marketing, and growth work. None of those paths assumed I would ever ship software.

Now I do. And it turns out the things I built because I was frustrated with a problem were the most useful things I made all year.

---

This article is part of an ongoing series on growth, AI-assisted work, and tools that actually solve real problems.