In the competitive world of content creation, small teams often find themselves in a David versus Goliath scenario, battling against larger news organizations with more resources at their disposal.
The challenge of producing high-quality, engaging content at a rapid pace is daunting and overwhelming, and the barrier to entry can seem insurmountable.
I say that as a member of a small, bootstrapped team that has been trying to solve the problem of how to only produce and distribute high-quality content with minimal time and resources. This is how we solved the problem.
Automating Content Creation
After accomplishing our first goal of being able to create content that we think is novel and unique, we started to share it on various social networks like Twitter, Pinterest, and LinkedIn. We were able to get some engagement here and there but soon realized that creating content started taking up all of our time which simply isn’t sustainable.
Tired and stressed, we were always trying to create the next great piece of content. Ultimately, we ended up building out our team. With several writers, a graphic designer, and an editor on board, we were ready to enter the ring.
Over the next several months, our content got better, and we were able to produce content faster. Unfortunately, the amount we were spending versus the ROI wasn’t working the way we had projected, and we soon realized that we needed a better solution, so we turned to AI.
🚀 CASI
After a fair amount of trial and error, we stumbled across an interesting discovery that we’re calling Collaborative Artificial Super Intelligence (CASI). To simplify, CASI is premised on decomposing intricate business processes into smaller sequential tasks and assigning different AI personas to each operation. So how does this apply to content creation?
As mentioned previously, our journey began as an ambitious endeavor to match, and even surpass, the content production quality and speed of larger organizations. We explored various AI tools to help us accomplish our goal, and after some trial and error, we identified the necessary functions for the tools we wanted to integrate.
1. Content Creation
2. Image Creation
3. News Aggregation
4. Content Distribution
After testing a multitude of different software, we arrived at a stack that worked great for our requirements.
The Stack
The culmination of this integration resulted in an open-source project QuasarAI, an AI-powered newsroom comprised of multiple AI personas, synergistically working that operates in the following manner:
And these are the tools that we ended up picking to make it all work.
- Feedly - Curates relevant news articles.
- OpenAI - Transforms these stories into captivating narratives.
- Midjourney and NextLeg API - Generate visuals to enhance the stories.
- StoryPRO - Manages the publishing of the final content.
- Aryshare - Oversees the distribution of content across various social media platforms.
Lastly, we needed to figure out an easy way to utilize prompts that interface with the AI personas to guide the tone of the content and the types of images being created by Midjourney. To accomplish this, we ended up making it possible to use a .yml file.
Getting Started
The deployment of QuasarAI marked a significant milestone for our team. We observed a substantial increase in content engagement, largely due to the system's ability to adapt its voice and style to resonate with diverse audiences. Setting up QuasarAI involves several steps and necessitates subscriptions to certain third-party services.
Here's a simple breakdown to get you started:
- Fork the QuasarAI repository.
- Duplicate the .env example file, rename it to .env, and fill it with the required keys and values.
- Generate and modify the blueprint files.
- Set up the database.
- Install Foreman and Redis.
- Launch the application and create an account with the necessary access privileges.
For a more comprehensive setup guide, please refer to Setup Guide and Deployment on the QuasarAI GitHub page.
Conclusion
The introduction of QuasarAI has been a game-changer for us. It has allowed us to automate a large portion of our content marketing with high-quality content that is on autopilot so that we can begin to compete effectively with larger news organizations for eyeballs.
We hope that by sharing our journey, we can inspire other startups and small teams to explore the transformative potential of AI in content creation.
References
Davenport, T. H., & Ronanki, R.. (2018). Artificial intelligence for the real world. Harvard Business Review..
Gatt, A., & Krahmer, E. . (2018). Survey of the State of the Art in Natural Language Generation: Core tasks, applications and evaluation. Journal of Artificial Intelligence Research.
García-Herranz, M., Moro, E., Cebrian, M., Christakis, N. A., & Fowler, J. H. . (2014). Using Friends as Sensors to Detect Global-Scale Contagious Outbreaks. PLOS ONE.
Guo, L., Tan, E., Chen, S., Zhang, X., & Zhao, Y. E. . (2008). The stretched exponential distribution of internet media access patterns. Proceedings of the twenty-seventh ACM symposium on Principles of distributed computing..
My Recommendations
Privacy Notice
We use cookies to store your preferences and analyze how StoryPRO is used.
Your settings can be changed, including withdrawing your consent, at any time on our privacy policy page.
Strictly Necessary Cookies
Marketing & Analytics Cookies
You agree to our use of strictly necessary cookies if you continue to use this website.
Content Restricted To Members
About Author