Published Nov 28, 2024
Perplexity as a Research Assistant
While working with early stage LLM startups over the past year, I’ve refined a workflow that integrates Perplexity API and Zapier to automate competitor research. As the field of AI is currently advancing at a rapid pace, keeping up with all the updates often felt burdensome. This system helped me offload the constant task of crawling the internet by generating weekly summaries of competitor updates and storing them in a dedicated Notion database.
The prompt is very simple - assuming that we’re interested in keeping on top of what news the leading LLM companies have posted recently:
Step 1: Setting Up the Perplexity API
Start by obtaining a Perplexity API key. With its single endpoint, “Chat Completions”, you can send POST requests that include your prompt.
The prompt is included in the “content” field with the ability to choose which model to use. The request looks like this:
Step 2: Automating with Zapier
Zapier handles the automation, running the following sequence weekly:
A scheduled trigger activates the process
Perplexity’s API processes the prompt
Results are forwarded to Notion for easy reference
The free plan offers a set number of interactions but for something relatively straightforward like this, it’s enough to get started. The setup streamlines monitoring competitor updates, offering insights without manual overhead.
In the example I’m using just a basic query but Perplexity’s API allows refining queries using additional parameters like:
search_domain_filter: Focus results on specific domains, e.g., competitor blogs.
Return_citations: Include references for deeper validation.