If I know exactly what I’m looking for, then yes, queries are good enough. But there’s still friction: logging into the DB, writing the query, running it, exporting results, plotting, then finally analyzing.
With AI, all of that happens in one place. It can query, interpret, and even generate charts if needed. And if I’m not sure what to search for, I can ask the AI to look at the data structure and suggest the best angles.
Excellent walk-through, Jenny! Actually working on a second brain post as well to show my workflow, and I was wondering of the second brain workflow of a couple people one of which was you! I was excited to go through this
Wow Jenny, every time I read one of your articles it opens my mind up to so many possibilities. This is so interesting and you’re right about the massive potential for MCP in the corporate environment. I sympathize with you on your trials with Copilot, facing the same issues in my experiments! Really great stuff here, thank you!
I’m glad the piece sparked ideas for you. The corporate side of MCP is wild… so much potential, but also so many roadblocks like Copilot. Makes me feel better knowing I’m not the only one bumping into those walls!
Pretty interesting!
I love the idea of connecting DBs to AI for interpretation.
Question: if you have the DB, aren't queries good enough? Why is the AI component adding value?
Thanks, Jose, really good question!
If I know exactly what I’m looking for, then yes, queries are good enough. But there’s still friction: logging into the DB, writing the query, running it, exporting results, plotting, then finally analyzing.
With AI, all of that happens in one place. It can query, interpret, and even generate charts if needed. And if I’m not sure what to search for, I can ask the AI to look at the data structure and suggest the best angles.
I guess that the queries are already baked in, right?
Because otherwise, the AI tool could try a different approach each time a prompt is received.
And additionally, how do you check if the response from the AI is accurate?
Love it. Definitely need to apply this model in my workflows.
So glad it resonated! Would love to hear when you try it in your workflow.
I like what you are doing! Thanks for the shoutout!
Happy to, you are sharing great guidelines! Thanks Kamil!
Excellent walk-through, Jenny! Actually working on a second brain post as well to show my workflow, and I was wondering of the second brain workflow of a couple people one of which was you! I was excited to go through this
Thanks a lot Joel! Love that you’re working on a second brain post, can’t wait to read it! 🔥
Wow Jenny, every time I read one of your articles it opens my mind up to so many possibilities. This is so interesting and you’re right about the massive potential for MCP in the corporate environment. I sympathize with you on your trials with Copilot, facing the same issues in my experiments! Really great stuff here, thank you!
Ahh, thank you so much Zain! That means a lot! 😊
I’m glad the piece sparked ideas for you. The corporate side of MCP is wild… so much potential, but also so many roadblocks like Copilot. Makes me feel better knowing I’m not the only one bumping into those walls!