MMarketing Against The Grain
← All episodes
06 November 2025

I Used Gemini Code Assist to Build a Newsletter App (for Free)

4Frameworks
8Insights

Listen

Frameworks in this episode

Marketing5 steps

Code-Powered Demand Capture

Turn useful interactive tools into audience and customer acquisition engines.

Read →
Innovation6 steps

Feature-Chunked AI App Building

Build and verify one functional slice at a time to keep AI coding agents aligned.

Read →
Productivity6 steps

Job-to-AI-Skill Prioritization

Convert recurring role responsibilities into ranked, reusable AI skills.

Read →
Productivity5 steps

Native-Expert Model Handoff

Use the model closest to a domain to brief the model executing the build.

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Hot Take· 2

Hot Take01:30

Interactive Tools Can Replace Static Lead Magnets

Traditional marketing funnels rely on content, forms, and subsequent conversion efforts. AI coding assistants make interactive tools and content nearly as easy to produce as PDFs, templates, and documents, creating a more useful entry point into the funnel.

  • Traditional funnels begin with passive content
  • Interactive tools provide immediate personalized value
  • AI has sharply reduced the cost of building software experiences
  • Marketers can use applications instead of static downloads

“What's happening now is these code and assistants mean it's as easy to create kind of interactive tools and interactive content as it was to…”

01:30
#lead magnets#marketing funnels#interactive content#ai tools
Hot Take09:00

The App-to-Newsletter Growth Model Has Flipped

Historically, software products added newsletters to support the growth of the application. Cheap, rapid AI-assisted development reverses that relationship: a purpose-built application can now serve primarily as an acquisition channel for a newsletter.

  • Apps can function as newsletter acquisition assets
  • AI reduces the cost of building narrow-purpose software
  • Code experiences can be created as quickly as content
  • Marketing teams can adopt product-led acquisition tactics

“In what world prior to AI would I have created an app to grow a newsletter, right?”

09:00

“But now it's been reversed because you can really create code experiences as quickly or even faster than you can create content.”

09:00
#newsletter growth#product-led growth#ai development#marketing strategy

Explainer· 1

Explainer00:00

Why Interactive Apps Are Becoming a Marketing Channel

AI coding assistants let marketers create interactive tools that capture demand and convert users into customers. These experiences can now be produced with effort comparable to creating conventional content assets.

  • Build useful applications that attract existing demand
  • Use interactive experiences to capture customer details
  • Convert tool users into an owned audience
  • Treat lightweight software as a marketing asset

“Using these tools to create code-powered experiences, apps that will capture demand that then you can convert into customers for your business is going to…”

00:00
#ai coding#audience growth#lead generation#interactive tools

Story· 2

Story03:30

Turning a Job Description Into Downloadable Claude Skills

The demonstrated app accepts a job description, identifies recurring tasks that could become Claude Skills, and generates the corresponding skill files. It transforms an educational topic relevant to the newsletter audience into a personalized utility.

  • Users upload or paste a job description
  • The app identifies suitable recurring skills
  • Each suggestion includes impact, effort, and rationale
  • Users can generate and download a skill file

“I can create an entire app to just allow you to upload your job. And that app in return will give you back those skills…”

05:00

“So now I can go to create and it creates the entire skill for me.”

07:30
#claude skills#app builder#job automation#lead magnet
Story08:00

Adding Email Capture at the Moment of Download

The host adds an overlay form that asks for an email address when a user downloads a generated skill file. Because the form appears after the application has demonstrated value, the tool can generate newsletter signups while still delivering a useful personalized asset.

  • Place the signup request at the download action
  • Let users experience the tool's value before gating the output
  • Connect the captured interest to a relevant newsletter
  • Market the application itself to generate subscribers

“What I would do when I'm kind of building this app is I would add a little form. And if you want to download that…”

08:00

“Now I can kind of market this app and I can create actual email signups for my newsletter.”

08:30
#email capture#newsletter growth#conversion#product-led marketing

Tool· 1

Tool02:00

Google AI Studio Connects Coding, Images, Voice, and Video

Google AI Studio's build environment can connect applications to Google's image, voice, and video capabilities. This integration makes it possible to create richer products, such as a conversational fitness coach, without separately assembling every underlying service.

  • Nano banana provides image capabilities
  • Gemini Lives API supports voice interactions
  • VO3 can add video or imagery workflows
  • Integrated services reduce app-building complexity

“The first thing you'll notice is it's actually integrated to Google's suite of tools, which it should not be underestimated.”

02:30

“If you want to create an app that's integrated with voice, it adds in Gemini Lives API, which I have done.”

02:30
#google ai studio#gemini#voice ai#multimodal apps

Takeaway· 2

Takeaway05:00

Use the Best-Informed AI to Prompt Your Coding Assistant

The host used Claude to prepare prompts for Google's coding assistant because Claude had stronger knowledge of its own Skills documentation and formatting requirements. The broader takeaway is that different assistants can complement one another rather than serving as mutually exclusive choices.

  • Match assistants to their strongest knowledge domains
  • Use one model to formulate instructions for another
  • Consult the source platform's documentation and guardrails
  • Build practical experience across multiple AI tools

“What I did for this, because it's Claude Skills, is I actually worked with Claude to actually build the prompts for Google because Claude would…”

05:00

“I could actually use one to prompt another assistant.”

05:30
#multi-model workflow#prompting#claude#gemini
Takeaway06:00

Build AI-Generated Apps One Feature at a Time

Even when a coding assistant can produce a complex application in one attempt, dividing the build into discrete features reduces confusion. The demonstrated sequence handled job-description parsing first, skill-file generation second, and downloading third.

  • Break an application into discrete functional parts
  • Implement the core input and parsing flow first
  • Add transformation or generation logic next
  • Finish with delivery features such as downloads

“The other thing is when you're creating an app, it is best to chunk it into parts so the assistant does not get too confused.”

06:00

“But to make sure this worked, I first of all built the functionality to upload a job or paste a description into this panel and…”

06:30
#ai coding#app development#feature planning#gemini