In English По-русски

Typical Startup Story on the Example of Fraza AI Telegram Bot

Published on September 22, 2023

Startups and Entrepreneurship

🇷🇺 This article is also available in Russian: Типичная история стартапа на примере Fraza AI Bot

One of the suite of my AI bots, Fraza AI Telegram bot is a quick experiment that I had built in just under 6 hours and deemed its launch a failure the same day.

Table of contents

A typical startup story, on the example of my Fraza AI telegram bot, had unfolded right before our eyes. I went from eagerly anticipating a rapid ascent to the crushing disappointment of shattered expectations... 😭 In less than a day, as opposed to a typical for startups half a year.

Yesterday, I posted an update in our work chat about rolling out a new version of a product I was building to staging. Doubtful about the grammar, I asked ChatGPT to rephrase my message to make it sound more natural, as I just had the OpenAI API playground open next to it. It worked. And I thought, "Ah, cool! I could quickly wrap it up into a Telegram bot."

After all, I have built the infrastructure for fast deployment of AI bots like that. 💪

It would be more convenient than going through a bunch of prompts. While pondering this, I couldn't help but let my imagination run wild. What if I could not only fix grammar and align it with American/British/international style, but also make it funny and, hopefully, worth sharing with others? Just imagine your message but written by a celebrity. Shakespeare, Tolkien, a character from Guy Ritchie's movies, Metallica, Eminem... the possibilities were endless!

In that moment, I could already feel the sparks of glory igniting! 🤡

"I'll roll out the bot, share it with friends first, then publish it for my subscribers, and finally write an article for a startup-orientated web-magazine..." — that's what crossed my mind. I quickly set up usage quotas and added convenient message editing, allowing the original text to be shared along with the corrected version.

"Sure, a Telegram bot might not be the optimal format," I thought, "but if people need it (and how could they not? It's freaking cool!), they'll definitely give it a try." Ideas of expanding to Slack, WhatsApp, and other platforms also danced in my mind.

So what's next? I published the post, which had been read by over 500 people in the first hours... but only three actually used it. Yes, my friend, only three! 🤣

There must be quite a few reasons for that

  • The text might have been poorly written, lacking a clear value proposition and a compelling reason for people to give it a shot.

  • Perhaps the target audience wasn't well represented among my blog subscribers, or the value proposition failed to reach them...

  • There were high transaction costs involved: you have to open the bot, click a button, copy your text, select another button (there were eight of them!), wait, read the response, and copy it back.

  • The perceived value fell short compared to ChatGPT and other editing tools, especially for individuals like my subscribers who aren't easily impressed with trivial things like this one!

Luckily, I had only invested around 6 hours in development. Not months or years. 😆

In the same series

Read next

Related projects