GPT assistants will be shocking for some of us.

For the past few months, we've been collecting conversations between bookers and clients, as well as agents and models.

Large language models (LLMs) are trained on a wide range of data, but each industry has its tone and jargon. Creating convincing emails that appear to be written by a human was not working so well using ChatGPT alone.

You may have come across articles either assisted or entirely composed by GenAI. These articles appeared somewhat generic, lacking a human touch. You wouldn't want your emails to be perceived in such a manner.

My strategy involved using a vector database, such as Pinecone, to store and manage the data. And I planned to fine-tune a GPT model with this data.

With the recent introduction of the GPT Assistant, I'm starting to grasp how it will once again change everything.

“An assistant is a purpose-built AI that has specific instructions, leverages extra knowledge, and can call models and tools to perform tasks.”

What I aimed to accomplish with a third-party vector database can be achieved using the Assistant's "retrieval" tool.

In this context, the additional knowledge refers to the collected questions and answers.

All I needed to do was give instructions to the assistant and upload a file with the prepared knowledge.

What comes next?

Another tool available to an assistant is "functions". In essence, the assistant can interact with other software. This could include accessing FAZE models "model-management "software or getting help checking travel options on BookingCom.

GPT assistants will transform the way companies view internal tools. Additionally, I look forward to seeing customer-facing assistants emerge in the coming weeks and months.

In your industry, how do you think assistants can be utilised?