You Don’t Have to Be a Developer to Build with AI: A Conversation with MetricFire GM Elliot Langston

Elliot Langston is not a developer. But with AI, he can turn technical product information into landing-page concepts, build microsites, and prototype personal app ideas far faster than before.

AI has helped Elliot shorten the path from idea to working concept, cut marketing-page deployment time by roughly a week and a half, and absorb much of the marketing work previously handled by a departing colleague.

We sat down with Elliot to learn how he uses AI, where it still falls short, and what he’d tell a beginner who wants to start building today.

Catch the full interview with Elliot inside Treehouse’s AI Bootcamp, launching August 2026. In four weeks of structured, project-based learning, you’ll build the working knowledge to use AI coding tools with confidence. It’s $25/month, self-paced, and built for people without a technical background who want to ship something real. Any plan gets you into the AI Bootcamp →

Meet Elliot Langston

“During COVID, I did take a two-year coding class. I have a key understanding of the engineering and coding concepts, but I would not say that this is a profession that I expect to transition to anytime soon.”

Elliot describes himself as a semi-technical beginner: enough grounding to prototype ideas, understand the basics of how products are built, and collaborate with designers and engineers when a project needs more advanced work.

“At this point, these days, all you need is just an idea.”

If you’ve ever felt somewhere between total beginner and professional developer, his workflow is a useful example of what AI can make possible.

How does a non-developer use AI at work?

Elliot’s most common workflow starts with work his technical team has already produced.

“Our team is building technical blogs about what MetricFire can achieve, for example, a new feature,” he said. “I’m thinking, okay, how can we turn this into a landing page and deploy this into something that’s marketable?”

The answer is Claude Code.

“I can use Claude Code, take that data, and turn it into a landing page. I already have the design assets for this, and then I have a landing page within one afternoon.”

From there, he works with the marketing, design, and engineering teams to refine and deploy the page.

He has also expanded beyond individual landing pages.

“Now it’s got to the extent where I’m able to create a microsite,” he said. “It doesn’t even have to be a landing page. Not just one page, but a page with different drop-down options for different use cases.”

Instead of starting with a Word document and handing off copy through multiple teams, Elliot can now begin with a working page concept that gives everyone something tangible to react to.

The numbers: what changed for the business

Elliot put a number on it.

Previously, a marketing-page project might begin with a Word document of copy, move to designers, then go to engineers, with revisions moving back and forth along the way.

Now, Elliot can create an early page concept himself, work with designers to refine it, and collaborate with engineers to push it live. The team also considers SEO requirements, including sitemap placement and domain structure.

“That has cut down deployment time by about a week and a half,” he said, describing it as a ballpark estimate.

AI has also changed how his role functions day to day.

“Before I [adopted] AI, I was working with a marketing colleague, and she decided to leave the organization,” he said. “Now with AI, I am fulfilling a lot of her role while still being able to focus on my organizational and sales responsibilities.”

For Elliot, AI has not replaced collaboration. It has made it easier to move from a marketing idea to a concrete, reviewable concept much earlier in the process.

Where AI still falls short

Elliot is direct about AI’s limits. He does not see it as a replacement for judgment, review, or creative expertise.

First, there is the human touch.

“I can just throw something out there, but it’s still not going to sound great,” he said. “You have to review anything before you push it live.”

For MetricFire, accuracy is essential. “Everything must be accurate to MetricFire’s functionality.”

AI can speed up a first draft, a prototype, or a concept. But human review, design judgment, and coding fundamentals still matter.

The mistake business owners make with AI

Elliot’s biggest frustration is tool bloat.

“I’ve seen a lot of AI tool bloat, where people are just using it for usage’s sake,” he said. “Vendors are pushing AI updates to justify increased pricing, even when it’s not valuable to what the actual tool is used for.”

The takeaway: do not add AI simply because a vendor labels a feature “AI-powered.” Use it where it removes a real bottleneck, shortens a slow process, or helps your team do work that was previously difficult to start.

What he had to unlearn

“I’ve been relearning and unlearning prompts so much,” Elliot said.

The tools change quickly, and he has found that prompts that work well in one product or model update may not work as well in another.

“I joined a Reddit community about key AI prompts, and it just keeps updating for ChatGPT and Claude Code with the most accurate prompts that I should and shouldn’t use on each update.”

He also keeps his own reference list. “I have a whole list of favorites for exactly what works best depending on whether I’m using ChatGPT, Claude Code, or Gemini for that specific project.”

That habit of treating prompting as a skill you keep refining, rather than a set of magic words you memorize once, is the same approach we teach in AI Bootcamp program. You build a system and keep improving it.

Two projects that show what is possible

Pin Studio: a personal app idea

“I couldn’t find any kind of pin collection cataloging app on the internet,” Elliot said.

So he started building one himself. Elliot collects pin badges and has a board with more than a thousand pins, and Pin Studio, which he is building with Claude Code in VS Code, is designed to help collectors organize them by themes such as Marvel, Star Wars, Disney, anime, Pokémon, games, and Japanese pop culture. It also includes different visual styles, including “Hero,” “Cosmic,” and “Retro.”

The project is still a work in progress. During the demo, Elliot noted that the photo-upload feature had a bug, and he was still working on login functionality.

His next challenge is learning the backend side of app development.

“Currently, I’m trying to learn how to host and create a database for an image hosting and account setup via DigitalOcean,” he said.

That is an honest picture of what it can look like to build with AI as a beginner: you can make meaningful progress quickly, but hosting, databases, user accounts, and file storage introduce a new set of technical challenges.

The two-hour microsite built for AI search

Elliot’s second project was built for MetricFire.

He wanted to improve the company’s discoverability in LLM-driven searches and Google results, so he created a glossary microsite based on the company’s sitemap and existing blog content. Each term has its own page with a definition, calls to action, and links to related blog articles.

“This was the key prompt, and right now I’m using Claude Code,” he said. “It gave me an initial idea here. We refined that and kept refining that regarding what kind of terms we want, what kind of logos to use, and the design.”

He built the initial version in about two hours.

“The final design looks pretty much the same as what we have on our homepage, just without the animation,” he said.

He stopped at 79 concepts because, beyond that point, he felt the terms became too vague. Afterward, Elliot and his team refined the project further for SEO and worked on integrating it into the existing site while paying close attention to its search impact.

That is a practical, replicable example of using AI to turn an existing content library into a more navigable and structured experience, and a strategy any marketing team can borrow.

How to start using AI at work, according to Elliot

Elliot’s advice is simple.

Start with an idea. “Just think up a concept, type the idea into Claude Code, and see what it spits out.”

Then refine it. “Refine it to what your imagination looks like.”

From there, you can learn the next layers: hosting, databases, monitoring, and getting real users into your application.

“At this point, these days, all you need is just an idea,” Elliot said.

For him, AI has made it easier to turn an early concept into something visible and testable. The next step is learning enough to build on that foundation.

A year from now, Elliot hopes to take MetricFire’s AI work further.

“I would like to incorporate an MCP server into MetricFire so that anyone could programmatically look into their metrics and dashboards just via AI.”

Want to build the way Elliot does?

You don’t need a computer science degree to begin building with AI. You need a guided path, a few fundamentals, and practice turning ideas into real projects.

That’s exactly what our AI bootcamp for beginners is built for: helping you move from an idea to a project you can share, even if you are starting with limited technical experience. It launches in August 2026 and is available with any subscription plan.

You can also browse our AI courses or follow the structured AI for Programmers track to learn how to use AI coding tools well. When you are ready to commit to a job-focused program, a Techdegree coding bootcamp gives you projects, feedback, and a portfolio.

Elliot summed up the appeal in one line: “It’s really helpful to cut down the time from project idea to realization.”

Start with an idea. Then use the tools, your judgment, and the people around you to bring it to life.

Want to learn more about MetricFire?

Elliot Langston and MetricFire on LinkedIn are two great places to start. Interested in giving MetricFire a spin? Sign up for a free trial and begin monitoring your infrastructure today. You can also book a demo and talk to the team directly about your monitoring needs.

Frequently asked questions

Can you use AI to build apps if you are not a developer?

Yes. Elliot Langston, a general manager with a basic coding background, uses Claude Code in VS Code to build landing-page concepts, microsites, and an early pin-collection app. He still relies on designers for visual polish and is continuing to learn backend, hosting, database, and account-management work with guidance from engineers.

What AI tools does Elliot Langston use at work?

He mainly uses Claude Code to build pages and prototypes. He also switches between ChatGPT, Claude Code, and Gemini depending on the project. Keeping a personal list of prompts that work well for each tool is a shortcut he uses, and he follows online communities to stay current as models evolve.

How much time can AI save a marketing team?

At MetricFire, Elliot estimates that using AI has reduced deployment time for new marketing pages by about a week and a half. Instead of beginning with copy in a Word document and moving through sequential handoffs, he can create an early working concept that designers and engineers can refine together.

Where does AI still fall short for business use?

AI-generated writing still needs human editing for tone and accuracy. Elliot also finds that AI-generated images and animation can feel artificial, so he works with a designer when visual work needs more polish. For marketing pages, he emphasizes that content must be accurate before it goes live.

What is the first step to start using AI at work?

Start with one concrete idea. Put it into an AI tool such as Claude Code, see what it produces, and refine the output toward your vision. From there, you can learn more about hosting, databases, monitoring, and how to turn a prototype into something people can use.