I recently wrote an article submitted to AAOA’s Rent Magazine (Q1 2024, p. 13) on using AI (Artificial Intelligence), like ChatGPT, as a virtual assistant to improve landlord-tenant communications. While researching for this article, I played a lot in ChatGPT around concepts of how best to use it to optimize short-form messages, like email. In that article, I was not able to use a great deal of the interesting outputs that ChatGPT was able to generate for me, but I wanted to make these examples available to the PropTech community and those we serve. This then serves as a collection of writing examples that a Landlord might generate with AI.
This Post is organized into 4 sections:
- Quick overview of what AI Chatbot is.
- Using ChatGPT to help a new user learn how to optimize its use. In other words, we should use AI to tell us how best to use the AI.
- Exploring different Tones of Voice would benefit landlords and PMs in communicating with tenants.
- Example output messages for two different communication scenarios:
- Notice to a tenant who is persistently late with rent payments and
- Informing a tenant about unexpected delays in resolving a maintenance issue.
- Direct Comparison of different ChatGPT outputs for the same message
A quick note to help you follow the pile of content in this posting. In the sections below:
- Text highlighted, bold, and italics are Prompts that I, as a user, typed into ChatGPT
- Within ChatGPT responses, text in block quotes is the output sample message from the AI.
Also, I precede all user input prompts with “User:” and all ChatGPT responses with “ChatGPT:“. So feel free to use your browser’s search function ([CNTL] F) to search for “User:” to just jump to each prompt to find ones of interest to you.
Very Short AI Chatbot Overview
(If you are not interested in what these AI tools are and how they are made, just skip this next quasi-technical bit.)
As a quick overview for those who don’t understand what AI-powered ChatBots are, let’s talk briefly about what makes them tick. This form of AI is based on two simple concepts:
- Natural Language Processing (i.e., the ability to parse and interpret natural human spoken and typed language) and,
- a learning system that allows the AI engine to break down a huge amount of content (we are talking about billions of pieces of text and writing) to find common patterns and concepts.
The AI is“trained” by allowing it access to a huge pile of content. It forms its own “understanding” of how the vocabulary in its language model relates to the concepts it sees in the writing. It then builds its own enormous database, defining how all of the stuff is related together.
This is not magic; it is just A LOT of brute force learning with some finessing by the programmers to optimize its capabilities.
(Part 2 releasing soon)
Read more in the next article in the series, “Using ChatGPT to Help Use ChatGPT“ >
