Let me tell you how I decided to prototype an idea, create a landing page, generate a logo and even start coding a MacOS app. I want to create it for myself so I figured I may as well release it. Will I? Perhaps!
Final result: https://meetingfusion.com
Word of caution - there are a lot of online grifters trying to sell you their secret sauce on how to make stuff (and money) with AI. The truth is you need to know a lot about what you want to do to get there. This article is about me sharing some of this knowledge.
Finding the right idea
I don't have any groundbreaking creativity tips for you, I just could use an app like this for my daily work and I think I'm not alone. That said, I have created startups in the past and I'm working at a cybersecurity startup Netacea managing teams of engineers and my role involves some product strategy development. This puts my mind in the right place.
Inspired by Pieter Levels' book MAKE, I also maintain a Kanban-like "Project Board". I put ideas there every time something promising comes to my mind. I then select the most promising ones or keep evolving ideas.
Columns in the board: Concept, Promising, Building, Success, Failure.
The idea for this app came from my daily work
I am in a lot of meetings
I prefer to focus on interactions not note-taking
That makes me forget and lose important knowledge
What if something took notes for me in a non-intrusive way
And made me always prepared as well as helped my teammates!
At the moment, I have a manual workflow consisting of a few separate tools for this. I will share it at some point, but let's focus on the product at the moment.
Because of security concerns, I can't use an app like this for a lot of meetings. Also, even though this could be considered equal to note-taking, if you're doing something involving recording someone, you may be legally obliged to inform them. And it's a nice thing to do anyway.
Working on the idea with ChatGPT
Brainstorming
Let's start with the initial prompt. Don't mind the grammar, I typed quickly.
Hey, I want to start a new product. I want to create a software for recording meetings.
One liner: Record your meetings, transcribe them, summarize and save to Notion.
Basically the app will use Mac screen recording api, then use whisper (locally perhaps) to transcribe meetings, then summarize them with get-4, get action points thanks to get and will sync all the contents with a Notion database.
The reason for using Mac screen record api is that when you record only audio on your computer, it usually just picks up microphone audio. Hence the audio quality is bad or it doesn't work on headphones.
In future releases it can contain ideas like:
Please help me expand on the idea and tell me what to do next. I want to start with some marketing and starting a waiting list for the software.
What I got back was:
I then focused on the last one, per ChatGPT:
Marketing and waiting list strategies:
Build a landing page: Create a visually appealing and informative landing page that highlights the features and benefits of your software. Include a sign-up form for users to join the waiting list.
Content marketing: Write blog posts, create videos, and share infographics about productivity, remote work, and meeting management to showcase the value of your software and attract potential users.
Social media presence: Establish a strong presence on social media platforms like Twitter, LinkedIn, and Facebook to share updates, engage with potential users, and gain followers.
Email marketing: Send regular updates to your waiting list subscribers about the development progress, new features, and expected launch date. Offer exclusive early access or discounts to build excitement.
Public relations: Reach out to journalists and influencers in the tech and productivity space to get coverage for your startup. Share press releases and participate in interviews or podcasts to spread the word.
Beta testing: Organize a closed beta testing phase and invite a selected group of users from the waiting list to try out the software, provide feedback, and help you improve the product.
Partnerships and integrations: Collaborate with other productivity tools and software companies to offer integrations that create additional value for users.
While I don't necessarily agree with every step, I decided to follow the advice and start building a landing page.
Let's start with Build a landing page:
Create a visually appealing and informative landing page that highlights the features and benefits of your software. Include a sign-up form for users to join the waiting list.
How do we do it? Be my assistant, tell me what to do, I will execute your ideas in the real world.
What I got back was a detailed 10-step action plan, which can be summarized as:
Choose a platform (Webflow, Wix, Squarespace, or WordPress with Elementor), pick a domain name (Namecheap, GoDaddy, or Google Domains), design your layout (headline, description, visual element, testimonials, CTA, social media links), write compelling copy, choose visuals consistent with your brand, optimize for mobile, set up a sign-up form (Mailchimp, ConvertKit, or Sendinblue), connect your domain, test and launch, and monitor performance using analytics tools.
I decided to use Carrd to create a landing page, I described how I use it in a previous article: read here.
I then followed with the next steps.
Choosing a domain name
I then turned to brainstorm domain names.
Ok, I already have a platform, I'm going to use carrd, I have a license and it has some prebuilt templates for such a landing page.
Let's go with 2. Pick a domain name. Can we work on it? Do you have any suggestions?
We iterated over 80 names in the subsequent messages.
I was creating shortlists of my favourite domain suggestions and used phrases and asked to generate ideas based on those.
I used bulk domain search to validate long lists of them https://www.namecheap.com/domains/registration/results/?type=beast
Purchased with Google Domains, which I now regret. It was a great tool but now was sold to Squarespace and I'll need to migrate. I've never used Namecheap, perhaps equally good, I feel a bit bad for not buying there.
I then asked friends about the 5 best ideas, and they voted and shared their feedback.
Ended up sticking with MeetingFusion.com!
Building a Landing Page
Carrd is the tool I know from building other landing pages. I knew I could build something extremely fast.
Copy - almost entirely with ChatGPT based on my ideas, got some extra ideas from friends. I followed the steps Chat outlined at the beginning. Then iterated hard and edited manually to make it sound better.
Icons - free for commercial use https://uxwing.com
Waiting list form - I asked ChatGPT which provider out of the available ones in Carrd to use. I settled with MailChimp, as I already had an account. Tutorial from Carrd.
Testimonials - after creating the prototype, I pitched the idea to some friends and asked them for testimonials
SEO - asked Chat: What page title, SEO description and meta tags to use?
Then I tested on mobile and other devices and made a few adjustments to font sizes, heading lengths and placement of elements
If you decide to use Carrd, you can use my referral code at checkout! 82DXK6L9
App Icon
I continued the chat by asking how do I go about creating a logo and favicon? We brainstormed a bit and I went on to use some of the AI tools I know.,
Used midjourney, just to start with, wasn't happy with the results.
I figured that perhaps the icon is not needed at this stage
I went to Bing Image Generator: modern minimal app for iphone icon with microphone, vector graphic - terrible!
I turned to Midjourney and Googled for some prompts people had success with: square with round edges mobile app logo design, flat vector app icon of a stylized "M" and "F" letters merged together, minimalistic, white background - well it didn't listen to the prompt but I got a nice result! Had to sloppily remove the background, it's not exactly square, still have some work to do! 🙄
Marketing the idea and next steps
ChatGPT suggested various marketing approaches, waiting list building and PR actions
And at the same time, I should be building the app based on my manual process
This post is knowledge sharing, but hey, it's a form of marketing as well!
I'll keep posting on Twitter (@horosin_)
Building the app
Well... I just typed in Can you generate Mac swift ui code for this app?
Guess what, after some back and forth - I got to a working UI. I had to learn some Swift (Mac programming language). With Chat's help, in the idea development chat and separately in a chat in which I started to ask questions about Swift, I came up with a working UI prototype in 2-3 hours, acquiring a new skill along the way.
This itself is a story for another article, let me know if you want to read it, I genuinely need some feedback!
Before actually starting to code, I watched these tutorials to get a general idea of what Mac app development looks like. I genuinely believe ChatGPT is great for horizontal skill development for senior devs.
Swift in 100 seconds: https://www.youtube.com/watch?v=nAchMctX4YA
General how to start coding an app: https://www.youtube.com/watch?v=6Qa-SspgRMM
Pinterest clone: https://www.youtube.com/watch?v=r0opgla9X-I&list=PLimqJDzPI-H_46mpSKnJgibfx3iUvYOvp
Menu bar apps:
https://blog.schurigeln.com/menu-bar-apps-swift-ui/
https://sarunw.com/posts/swiftui-menu-bar-app/
ChatGPT was oblivious to some new features and SDKs so I had to copy some documentation into it. I've also asked Bing Chat to do research and pasted it into ChatGPT. That was before I eventually got access to web browsing in ChatGPT.
Conclusion and what next?
Well, this was a lot of fun!
The process was great, GPT significantly helped me with writing copy for the landing page. Brainstorming bits were also great. I loved the domain ideation, it was one of the worst things about starting a project in the past. Time sink with little benefit. But I always had to have a name before diving in.
That said, I still would have gone nowhere without people I could ask for opinions and get real criticism from. I also wouldn’t have created a landing page or started coding at all if my mentor didn't dare me to do it a week before!
The cover photo was generated with Bing Image and extended with Dall-E.
