SUBSCRIBE | ADVERTISE

Vibe coding my own sales tool

It was going to be a weekend project. Two months later, I have custom sales tool and a junior rep named Codex.

Was this email forwarded to you? Sign up here.

TLDR; I vibe coded a tool for sales/outreach. It was supposed to be quick but ended up taking me 2 full months. Owning the infrastructure means I can now spin up new features fast, and having Codex wired into my database and email is a surprising perk.

In this email:

• Why I needed this

• Building a sales tool that works for me

• I gained a lot, but didn’t save much

• Should you build your own?

If you’re flirting with the idea of vibe coding your own tools, this is for you.

The why: I was done with spreadsheets and manual research

I have some awesome sponsors in this newsletter (shout out to today’s sponsor Evermind).

I’ve been lucky so far in that sponsor inquiries have been mostly inbound, in other words sponsors reaching out to me. But I also do outreach, contacting companies that I want to have as sponsors.

As any creator will tell you, communicating with potential sponsors and making deals, whatever form that takes, has a lot of moving parts: companies, decision makers, proposals, follow-ups, campaign formats, creative assets, invoices…

Previously, I was juggling spreadsheets, email accounts, templates and spending time on manual research to find potential companies to contact. That’s not really where I want to spend my time.

I was looking for a way to streamline the process of finding relevant sponsors, reaching out to them in a way that’s effective, making deals and coordinate the deliverables.

My goal was to create something that allows me to spend as little time as possible selling and coordinating sponsorships, so I can focus my time on building new things with AI and creating interesting content for you.

IN PARTNERSHIP WITH EVERMIND

EverMemOS — The next-generation AI memory system

Most AI agents forget everything after a session—making them inconsistent, hard to debug, and impossible to scale.

Inspired by human brain memory mechanisms, EverMemOS provides an open-source memory OS that supports 1-on-1 conversation scenarios and complex multi-agent workflows. As reported, EverMemOS achieved 92.4% on LoCoMo and 82% on LongMemEval-S, both SOTA results of the two benchmarks.

Later this year, EverMind will launch the cloud service version, offering enterprise users advanced technical support, persistent storage, and scalable infrastructure.

If you're building agentic apps, EverMemOS gives you the memory layer you’ve been missing.

👉Explore EverMemOS

Blogs

Building a sales tool that works for me

I decided to spend “a few days” vibe coding my own sales tool in Cursor, thinking it would be fairly easy.

After a few days I had a really nice design and an app that worked… about 80% of the way.

From experience, I know the last 20% always takes 80% of the time. I sometimes think my brain conveniently “forgets” that to get me started.

Getting this tool to work at a level where I’m comfortable using it ended up taking two months of full time work. Yikes.

I started using the tool last week, and have closed one sale with it so far.

Here’s what the tool actually does:

Finding and enriching leads

The tool helps me find potential customers and make them ready-to-contact:

• Scrapes and imports names of companies that have advertised in other newsletters recently (warm leads).

• Uses AI to enrich leads with a summary about what the companies do and classify them into industries so I can easily filter strong matches for my audience.

• Enriches companies with a list of employees and their emails, then uses AI to guess the best decision makers to contact.

• Doubles as my CRM. I store data on previous sponsors in this list as well.

Outreach and personalisation

The tool helps me send out tailored sequences of emails to my prospects:

• Gives me a unified inbox where I can connect any number of e-mail accounts, and send/receive emails from the app. Messages are conveniently aggregated by company

• Allows me to create custom templates for outreach/follow-up, including information I have enriched the leads with.

• Lets me build tailored sequences of templates to customize the messaging flow and A/B test.

• Gives me basic stats on sends, opens, replies, and clicks.

Codex is my new sales rep

This part surprised me.

Having built this myself, I realised that I now have an AI agent that can read and take action in my database and email accounts.

For example, I can ask Codex to research my database of leads, shortlist the best-fitting companies for my next newsletter topic, then use a template to send a personalised email to each decision maker in those companies. All with a single prompt, and I can keep track of what has been done inside the app UI.

It feels a bit like having a sales intern who’s always up for the job.

I gained a lot, but didn’t save much

Building this took longer than I had imagined. Luckily, some good things came out of it.

New features are easy to build

The coolest benefit is that my little business essentially has its own super-custom infrastructure for managing sales.

This lets me build new features and tools later (scheduler, invoicing, etc.) without starting from scratch each time.

Let’s say I need a new type of AI enrichment (e.g. what’s the company’s total funding); I can go into Cursor and probably spin it up in a couple of hours. I won’t have to pray that some third-party tool has it, and pay for that additionally.

The cost savings are meh at best

The good thing first: I’ve essentially replaced paid services I would have needed without this tool.

A CRM like HubSpot is often $15-50/mo. Lead enrichment with something like Clay, Instantly or Lemlist? Easily $100+/mo. And I’d probably only need a fraction of their features.

My tool costs only ≈$50/mo to run (APIs, database, email lookups) with lots of usage. And it has an agent inside.

That said, I could probably have been happy with an existing tool. And if I put a price tag on the time I spent building, the math is pretty clear: I didn’t save any costs on building.

Learning is the real ROI

The biggest win for me personally is the experience I’ve gained. Knowing I can design and ship systems that solve real and messy problems, end-to-end. That feels empowering.

And I’m sure the next thing I make will be quicker and easier, both because I’m more experienced as well as the AI models getting better by the day.

The learning experience of shipping this was probably more valuable to me than the mere utility of the tool.

Should you build your own?

I recommend vibe coding your own tool only if you value the learning experience greatly. It takes a lot of time, effort, and can be frustrating at times.

I don’t recommend vibe coding a personal tool to save costs, unless it’s a very simple app. The time it takes to ship something complex will probably not be easy to justify in dollars.

The type of scenario where I’d personally consider vibe coding my own tool again is a workflow that’s both very important and very specific to me; something that pre-existing tools don’t really handle well.

And remember—creating something truly useful still takes real finesse and lots of time, no matter if it’s you or AI typing the code.

THAT’S ALL FOR THIS WEEK

Have you vibe coded apps to automate parts of your work? Hit reply and let me know!

Affiliate disclosure: To cover the cost of my email software and the time I spend writing this newsletter, I sometimes link to products and other newsletters. Please assume these are affiliate links. If you choose to subscribe to a newsletter or buy a product through any of my links then THANK YOU – it will make it possible for me to continue to do this.