AI in Procurement - My first steps with Prompting

Over the last year, I’ve dived into the fascinating world of AI, aiming to unlock new doors in procurement with this technology. My adventure started from a simple curiosity sparked by the buzz around ChatGPT, with a strong focus on already existing applications. Now, it’s evolved into an exciting private project: building a simple AI Assistant for Negotiation Support.

When I first started, I thought using AI tools like OpenAI would be straightforward. But recently I learned that talking to AI, or 'prompting,' is much more than just a chat. It's a blend of art and science, similar to coding but with words.It was a bit like trying to bake a cake without a recipe at first – a lot of trial and error. In the end, I decided to scrap the first version altogether. 

This led me to take a step back and learn more about how to effectively communicate with AI. I attended a short course on prompting (from IBM), which was the right decision. It opened my eyes to a much wider view, showing on the horizon possibilities for using AI in procurement (automating documents, negotiation processes support…).

But this step of my journey has taught me more than just the technical skills of working with AI. Prompting could very well become a fundamental skill, much like coding was when digital technologies first emerged - but for everyone. 

If someone were to ask me: Can you "program" a simple AI Assistant purely with prompting? My answer would be: YES. Of course, there are many gradations and complexities. On the other hand, there are many tools that make this possible even for non-tech people like me. 

Which tools do I use at the moment:

- GPT Builder from OpenAI

- Playground on the OpenAI platform

- MaxAI to get even better prompts and to write instructions. Offers a few different LLM models.

Sources to help me get the knowledge:

- https://skills.yourlearning.ibm.com/

- https://www.promptingguide.ai/

 - https://platform.openai.com/docs/overview

- https://www.datacamp.com/tutorial/open-ai-assistants-api-tutorial

- https://www.linkedin.com/pulse/how-build-gpt-based-ai-assistant-insights-lessons-learned-majewski/?utm_source=share&utm_medium=member_android&utm_campaign=share_via

- https://www.pluralsight.com/resources/blog/data/openai-assistants#playground-ui

- https://tecnosoluciones.com/how-to-create-an-artificial-intelligence-assistant-without-being-an-expert/?lang=en#What_steps_should_be_followed_to_create_a_Generative_Artificial_Intelligence_Assistant

This is just a small selection of generally accessible sources of knowledge. 

In the next few posts I will tell you a bit more about my approach to the project. I am therefore extremely open to comments, ideas, tricks and hints. 

What about you? Have you started exploring AI in your field? How do you see it changing the way we work?