As the bullet train called Generative AI moves along we can only stop and marvel at its progress for a short while. That is, until the next big announcement hits us in the face, right? I think it’s a good idea to “niche down” and not look at everything the GenAI ecosystem has to offer. It’s time to only look at the thing that interests you the most, and find the GPT / AI companions that can assist you with that.
For me, I want to help the PMO succeed with Microsoft Technology. This has led me to create a couple of ChatGPT models. Let me introduce them.
Note: I’ve been away for a bit because I was moving the family to our new house. Seems these GPT models are now available and free for everyone to use, and that you only need to log in. Please let me know if I misread that message.
![](https://www.theprojectcornerblog.com/wp-content/uploads/2024/03/oil-painting.webp)
Yes, Oh Yes! That is definitely me!
What are Custom GPTs again?
As I’ve described in a previous post, you can now build a GPT that follows specific rules and has a tendency to focus on specific topics.
You can give the GPT a tone of voice, make sure it describes in large detail or just give you a quick answer and so on.
The custom GPTs are a good first stop when it comes to considering building a dedicated chat bot in any company. Sort of, getting your feet wet before diving in a more complex system such as Azure OpenAI.
I’ve created 4 GPTs up till today. They were made because I wanted to find out how to do this, but also because of a personal need for them. Vicky especially is a resource I regularly consult while building custom Deneb visuals.
Here are the different GPTs I’ve made so far, the links can only be accessed with a account to ChatGPT:
The first GPT model: The PMO Pro
This very first GPT model didn’t have a “name” like the rest. But its current 700+ users don’t seem to mind.
The model was designed around a prompt that mimics a professional PMO manager. You can ask it any Project management related question and it will respond in a friendly and educational manner. The end goal (very similar to my own) is to help the PMO succeed.
Access the GPT here: https://chatgpt.com/g/g-9eBBn6lcn-the-pmo-pro
Vicky Vega, my personal Deneb & Vega tutor
Before Vicky I’ve tried to learn Vega by accessing YouTube tutorials and reading the extensive documentation. But because either they were too broad or in a completely other niche than my own, most of the time I didn’t get far.
Besides, creating Deneb visuals isn’t my core competence. And I wouldn’t be able to allocate large amounts of time to increase my knowledge.
With that thought in mind I set out to build Vicky Vega, a cool name I thought and one that would stick in the minds of people seeking its help. The original profile picture design was a “biker chick blowing bubble gum, 80s style”.
Vicky was instructed on the official Vega and Vega light documentation and Deneb-specific resources. As such, she has all the core knowledge to be able to fix any Vega-related code we throw at it.
Just like the other GPT, I made sure this one wasn’t only friendly it needed to be educational as well. I think of Vicky as my personal tutor on the Vega and Deneb journey. The 800+ users of the GPT might agree with this thought. It’s been featured on the Deneb website and it is part of presentations by Power BI instructors already. So proud of my girl 😁.
Access Vicky here: https://chatgpt.com/g/g-JxECNXqsw-vicky-vega
![Vicky Vega, my Vega GPT model designed to assist me with all my Deneb visual creation.](https://www.theprojectcornerblog.com/wp-content/uploads/2024/06/image-1024x342.png)
Dave Dax, my DAX sparring partner (until Power BI has Copilot)
In the line of thought for the Vicky Vega GPT model, I decided to do the same for DAX queries in Power BI.
Dave is set to assist a user in developing measures and columns for DAX inside Power BI specifically. And he’s supposed to help out the user in an educated manner.
The model will likely be the first one to become obsolete. I’m basing this assumption on the fact that Microsoft is so heavily involved in Gen AI at the moment, that I can’t imagine a situation where one of its flagship applications won’t have the capability to deliver its own Gen AI DAX suggestions.
An internal Gen AI solution like Copilot would be very capable of assisting us. Once the solutions mature beyond taking in more than a single table (like Excel at the moment). Just imagine if we could feed the model to Copilot and it would know what type of chart we want and where. The “analyze my data” feature hasn’t been very good at it’s job, I think.
So, until Power BI ramps up enough, join the 50+ users and access Dave here: https://chatgpt.com/g/g-vfcWrMODh-dave-dax
Nelly “PMO” News, getting the latest PMO Scoop
Nelly is a completely different beasty. While the previous GPTs were mainly focused on assissing or solving niche topics, Nelly takes on a broader approach.
Just like GPTs such as Concencus and the Innovative Research Explorer GPT, Nelly looks at the latest news regarding a specific topic.
But where the previous two focus on research papers, Nelly focuses on Project Management related news.
This is the new kid on the block, and a first attempt in line with these other news gathering GPTs. Seeing as there have only been 4 users of the model so far. I don’t think it’s a big success. Maybe you want to try it out and tell me what’s wrong with it?
Nelly can be found here: https://chatgpt.com/g/g-bIyuklBRr-nelly-pmo-news
Final notes regarding GPT Models
This is my first article since “the big move”. The family and me are finally settling in in our new house in Groningen (Northern region of The Netherlands). Generative AI has come a long way since I started the series in March 2023. I can’t wait to see what the next big thing will be.
Now that I’m back, you can expect to see new videos coming to the YouTube channel as well! I’m continuing, at least I hope so, my 2 videos per month schedule.
And JUST before you go, there’s also a newsletter you can subscribe to. It’s free and you get access to files I used in my videos and blogs.
Happy to be back,
Erik v H.