ABOUT
The AspIyr Pitch Deck is another brand inspired project. This time, the goal was to create an Automated AI Agent that can take a simple input and produce a nuanced output with actionable insights that users will find highly valuable and useful.
The tools and software used in this project include Typeform, n8n, Google Docs and LLM's like ChatGPT and Perplexity. The final result is a Mock Pitch Deck PDF that can be used to further research a business idea and potentially develop it into a full-scale venture.
WHY?
As an aspiring entrepreneur, it only felt natural to have one of the projects be related to the process of venture creation, be it the ideation phase or the development of a Minimum Viable Product (MVP).
On the basis of a vague vision and a brand name, I created the advertisement. However, I wanted something more tangible that would further solidify the vision for not only myself, but all of you who are currently reading this and might be considering tagging along on this journey!
The decision to use an AI Agent to achieve this is based on a little theory I came up with, which is that we are currently in Phase 2 of the AI Evolution - Agentic AI. Phase 1 was LLM's and Generation models. Phase 3 will be the birth of Artificial General Intelligence(AGI). Therefore, it is absolutely essential for anyone working with AI to be capable of creating AI agents and Automated workflows.
HOW?
PHASE 1 - Choosing the Automation Platform & Input Method
Just as the Mitochondria is the powerhouse of the cell, the automation platform is what makes this project possible to begin with. Therefore, choosing the right one was an important decision, and I decided to go ahead with N8N. While it has a steeper learning curve and is more expensive when compared with other options, it also has better long term scalability as a result. As a budding entrepreneur, it was a great choice.
As for the Input method, TypeForm was a no-brainer, given that form submissions are still the most commonly used methods of gathering data. TypeForm makes the process more interactive and enjoyable (NOT SPONSORED BTW)
PHASE 2 - Building the Workflow
As this was my first attempt at building an agent, I required guidance from someone who could explain it to me like I am 10 years old, and that's how I stumbled upon this brilliant Tutorial by Tina Huang - VIDEO. The core elements (nodes) in the workflow are as follows -
The INPUT node: This is what triggers the workflow, i.e. a Typeform submission.
The AI AGENT node: This is where the orchestration happens. It combines multiple sub-nodes to transform raw user input into a structured, professional pitch deck. Within this agent are the following components:
(i) Chat Model –
The Chat Model (OpenAI) is the language engine that rewrites and polishes the raw inputs from the Typeform. It ensures the user’s ideas are expressed in a clear, professional, business-ready tone.(ii) Memory –
The Memory node stores context across the workflow. This ensures that information gathered at the start (like the user’s original idea or selected industry) remains accessible to the Chat Model throughout the process.(iii) Tool (Perplexity) –
Whenever the Chat Model needs supporting data (industry insights, market statistics, trends, references), it calls this tool.The Perplexity node returns real-world, source-linked information, which the Chat Model integrates into the pitch deck.
The OUTPUT node: Finally, the structured output is sent to Google Docs. Here, placeholders in a pre-designed template are replaced with the generated content. The result is a polished, professional PDF that is then sent to the user at the email address provided in the TypeForm.
Result
The outcome of this workflow is seamless end-to-end automation: a user submits their idea via Typeform, the AI Agent rewrites and enriches it with research, and the content is instantly formatted into a mock pitch deck, sent directly as a PDF to the user’s inbox, ready to share or present.
The attached screen recording captures this process in action, showing how quickly an initial idea becomes a tangible, high-quality output with minimal effort required!.