Peppertype.ai

The Story Of How We Built Peppertype.ai In 15 Days!šŸš€

Rishabh Shekhar
ā€¢
Posted on 23/03/21ā€¢9 min read
The Story Of How We Built Peppertype.ai In 15 Days!šŸš€
ā€¦.and the onset of the Pepperā€™s ā€œContent Stackā€

What is Peppertype.ai?

Peppertype.ai is an AI assistive tool built on top of GPT-3, an autoregressive language model that uses deep learning. We discovered that this technology had the potential to help businesses and creators generate content ideas instantly and also create short-form content copies.

Imagine this as a fountain of never-ending creative ideas, taglines, descriptions, bios, and all other forms of copywriting that you can tap into whenever you wish.

 

And why is it important?

It is a step towards us realizing our dream of a full-stack content creation platform. It brings us that much closer to our collective vision of making the process of content creation more innovative and seamless!

At Pepper, we have always envisioned a content stack, just like one for Product Management (Notion, Heap, Loom, Jira, Confluence, Heap, Retool, Metabase, Hotjar). Peppertype.ai will be the first of many in the stack that we aim to build for our friends, spearheading content as creators or businesses.

Since Pepper’s inception, our theory was that the content creation experience should be based on Brain Chesky’s (Co-Founder and CEO of Airbnb) famous analogy of how Airbnb thrives not just to be a 5-Star experience but to be an 11-star one! A 5-star experience in the content will be you giving a brief and the writer creating the content exactly how you need it. But we donā€™t want to stop there! We want to help you create briefs that deliver amazing content, devise a winning content strategy, and much much more.

We donā€™t want to do the best by industry standards. We want to create industry standards.

Letā€™s start from the beginning.

Building a startup in any space opens you up to so many different problems that plague that particular industry, and starting up Pepper was no different. Back in 2017, our problem statement was simple when it all began: solve all problems with content creation. Except it wasnā€™t as simple as that! The content creation process has many parts to it, and every aspect needed some fixing.

From a businessā€™s point of view, if it wants to create content at scale, the first problem is a lack of ideas. What content to create? Thatā€™s the question. If thereā€™s one thing that our experience with over 600 businesses, agencies, and individuals on the Pepper platform has taught us, the struggle to create short-form content like ad copies, etc., is real. Moreover, it slows down the entire content cycle. Pepper could step in to create that content, but what about when they are stuck at ideation itself? How could we simplify and speed up the manual and exhaustive process of generating content ideas?

The answer came to us in a rush. To democratize content creation and give the power back to brands and creators, we needed to make the atomic pieces of content more accessible. We also needed to integrate creativity into the content creation process and solve for it.

How could we, as Pepper, sitting at the cusp of this massive repository of insights from businesses and creators, leverage all of this to solve these problems using technology?

Peppertype.ai: The Origin Story

This was the big question we were discussing in one of our brainstorming sessions back in December 2020. In the post-pandemic world, discussions had become more virtual. Still, luckily at Pepper, we managed to retain the essence of physical meetings as seen in every early-stage startup: a freewheeling conversation accompanied with a lot of laughter and casual exchange of ideas. Add to that a few instances of call drops, frozen screens, and voice lags, and youā€™ve got the ingredients for a successful virtual meeting.

Someone enthusiastically started to talk about OpenAI, the worldā€™s leading organization in AI research, and the GPT-3 model that they had released in July 2020. We had all been watching these developments closely, partly because weā€™re all tech enthusiasts and partly because some of us could sense that this technology could disrupt the content marketing world as we know it – something that is almost an unsaid mission statement at Pepper. GPT-3 had been trained on 175 billion parameters and made language creation easier and much more accessible. The reviews that it was receiving were terrific. It was fascinating, but it needed a bit of spice to it, or as we may say, a bit of Pepper.

During our ideas exchange, we realized a clear use case here for businesses as well as creators. For businesses, it is how do they get topics created for their blogs? For creators, it is the headline of a webpage that they are working on. For general creators like us, it was an idea which we could use to post on our social media handles!

It was time to put our thesis to test. We had access to GPT-3 already, and we started toying with the idea of how we can build a tool on top of the algorithm to solve such issues. We did a pilot of talking to 20 creators and few businesses and got a brief idea of whether they would pay for it or not – and we were blown away by the excitement folks showed for such a product. Peppertype.ai had to be created, and it had to be done by us.

The first glimpse of validation

ā€œIf at first, the idea is not absurd, then there is no hope for it.ā€ ~ Albert Einstein

A tool that can create content on its own? It sounds like magic, doesnā€™t it? Creativity and imagination arenā€™t available on tap, nor is there an infinite fount of creativity somewhere (trust me, Iā€™ve looked). Even the best of authors and writers struggle with writerā€™s block. It makes the process of content creation inefficient and time-consuming. And Peppertype.ai could make it all better.

We quickly whipped up an MVP and visited the 20 folks we had talked to earlier and got them to use it in front of us. Some of the reactions we got from the incredible people at Pepper Content blew our minds:

ā€œThis works like a miracle!ā€

ā€œARE YOU TELLING ME I DONā€™T NEED TO SPEND HOURS STUCK AT A POINT AND IT CAN BE DONE IN A MATTER OF SECONDS?ā€

ā€œLove how smart this is! Feels unreal!ā€

Oh, the excitement of getting an overwhelming initial response and validation, even before the product is launched! We now knew we were on to something big and the positive feedback was the fuel we needed to go out all guns blazing. The only thing left to do now was to create this world-class product. Easy.

 

Rome wasnā€™t built in a day, but Peppertype.ai was built in 15.

ā€œOnly move forward with creating a product that will be ā€˜above the bar.ā€™ā€ ā€“ Brian Lawley, CEO, and founder of the 280 Group

We decided to build Peppertype.ai on the existing main Pepper product stack, making it easier for us to reuse some portions and elements of the frontend and just plug and play! We also decided to use the same payments infrastructure that we were building for the Pepper Platform. A famous saying goes, ā€˜Rome wasnā€™t built in a dayā€™ – and an addition to that says, ā€˜but they were laying bricks every hour.ā€™ And that is the power of having a long-term vision. Our super agile, small team of 3 developers started to build the foundations of a stellar product over the bricks we had laid since the birth of Pepper. We were strict on our iterations; daily catch-ups were mandatory, and we eventually managed to pump out the first version of the product in 15 days.

The project was memorable as someone took out time for us to help build this out. I would like to specifically mention how Nirant K, ML engineer at Verloop.io and a senior from BITS Pilani, was monumental in helping us figure out how to build upon GPT-3. He brought his immense expertise in NLP and Machine Learning guided us on our engineering journey like the tech sherpa he is.

The complete application was developed using Javascript primarily as the programming language, with the ReactJS framework used for frontend development and NodeJS for Backend development. Weā€™ve extensively used AWS for hosting and running the application, which enabled quick prototyping of the application. We were committed to the KISS (Keep It Simple, Silly) mentality and avoided over-engineering, which saved the team many late nights (and a lot of coffee). And for the final touch, we used Stripe for managing subscriptions, which is a proven solution, so we didnā€™t need to reinvent the wheel. We were hoping to use Razorpay for our platform, but due to some issues on the compliances side, we had to plug in Stripe at the last moment. Key takeaway? Always build a payment system that can plug in any payment gateway. Reduces your reliance on any one tool in particular.

Getting the design right

ā€œOnly when the design fails does it draw attention to itself; when it succeeds, itā€™s invisible.ā€ – John D Derry.

We wanted to stick to a straightforward and sleek design. We wanted our users to understand how to generate content in seconds with very minimum inputs.

When a user lands on the page. We wanted their journey to map to three elementary steps:

  1. Choose the kind of content you want to create (Twitter ideas/blog ideas)- Which industry would it cater to, what type of content etc
  2. Give us the name of your brand and the essence of the product in 180 characters
  3. Bookmark your favorites from the AMAZINGGGG content ideas generated in seconds.

We worked on the evergreen stack: Adobe XD (to create wireframes and high-fidelity designs), Google Analytics and Hotjar (to analyze user behavior), Notion (to list down tasks and Todos), Slack and Google Meet (team collaboration and all the discussions).

The day of the launch

When you launch in a rocket, you’re not really flying that rocket. You’re just sort of hanging on – Michael P Anderson.

Post a rigorous 15 to 20 days of building and testing; the product was live for an internal launch. Rohit, who heads Product Management for Peppertype.ai, presented a demo of the product to an almost clueless crowd. And they loved it. Many business folks in our team saw massive potential in how this would make life easier for them and the clients. But the most common feedback was that it looked beautiful.

One of the most important thing for a person marketing a product is to believe very strongly in it. “You are your first User”.

Fun fact- all the content for Peppertype.ai was written using Peppertype.ai and we got some shoutouts for the same.

Funny incident, during a live demo, Pranav tried out generating captions for Peppertype.ai itself and one of the output said ‘The Future Is Here and It Says “I will write for you”‘ which is beautiful.

How we managed the launch on Notion

The launch was spearheaded by Pranav, a junior from BITS Pilani, and he nailed it! Our launch strategy was simple:

  1. Announce a waitlist to see the initial response
  2. Build up enthusiasm around the product through carefully planned content. A cohesive content calendar across all platforms is a must.
  3. Integrate all communication channels to ensure maximum outreach. We activated all our distribution channels, be it emails, social media, or WhatsApp.
  4. As a content tool that boosts creativity, we had to get creative in our messaging. So one by one, we started generating blog ideas, Tweet copies for brands we love – from Zomato to Apple to Durex.
  5. Go demo-crazy. One other thing that made a difference for us was talking to folks and giving them demos – as many as we could! The personal touch of explaining a product and what it can do for you can never go out of style. This led to a lot of word of mouth and even more sign-ups! I think launching a product means having the guts to do cold outreach and hope that the person agrees to see your product.

The response

I think one of the most significant validation came when Ryan Hoover, founder of Product Hunt himself, loved the product design! We also received a lot of love on Twitter from tech enthusiasts and creators alike.

24th February: Peppertype.ai goes live

The soft launch’s tremendous success gave us the confidence to go all out, and on 1 March 2021, we announced our pricing models.

The buzz has anything but died down. With over 30 media mentions and 35% of users from outside India with almost zero ad spending, we feel good. Within four days of launch, we crossed $500 and aim to take this to $1 Million ARR very soon!

ā€œDonā€™t forget to have fun.ā€

..it indeed, was fun.

The time when Google generated a tagline for us šŸ˜†

We think that a product is a result of the collective experiences of everyone that is involved in building it, and thus, it is almost imperative for everyone to have fun on the way. And we sure did!

As we build Peppertype.ai plus the platform, the ultimate vision is to help you strategize creation while enabling a one-click order flow from the same screen for content on the topics you like! Imagine how seamless things can be; if post creating blog topics, you can click on ā€œOrder This Content,ā€ and a kickass creator on our platform immediately gets to work on your blog piece. Wouldnā€™t that be just spectacular!

Watch Peppertype.ai in action!

And so we move towards a future where content and technology work together seamlessly to aid businesses and content creators all over the world. Weā€™re proud to have created a product for the world, from right here in Mumbai and the remote team spread across the country. Whatā€™s most exciting is that here at Pepper, innovation hasnā€™t stopped during the pandemic. It has only accelerated and motivated us to do more and build more inclusively for the content ecosystem.

 

Onwards and upwards!šŸš€