Please don’t put your life in the hands of a rock ‘n’ roll band, who’ll throw it all away. – Noel Gallagher
Noel Gallagher probably wasn’t talking about AI. But he could have been.
General-purpose AI is extraordinarily powerful. ChatGPT and Claude can read documents, analyze financial information, build dashboards, and tackle increasingly complex work.
So we hear the question all the time:
Why do I need Notessa when I already have ChatGPT or Claude?
Because there is a difference between AI that is capable of doing commercial real estate work and AI that is built to do it.
Stop training your AI
Think of a general-purpose LLM as a brilliant new employee who has never worked in commercial real estate.
Before they can really help, you have to teach them the business: the terminology, the documents, how transactions are structured, what matters, and how you want the finished work presented.
Now imagine having to teach that employee the same things over and over again.
That is essentially what happens when users repeatedly write detailed prompts, add instructions, correct interpretations, and refine outputs.
It’s time-consuming. And saving time is supposed to be the point of AI.
Notessa starts somewhere different. CRE knowledge, document intelligence, and workflows are built into the platform. You shouldn’t have to teach your AI how to do the job every time you give it an assignment.
Same documents. Same assignment. Very different answers.
We decided to put that idea to the test.
We gave Notessa and Claude the exact same set of transaction documents and the exact same assignment:
Create a capital structure for the transaction.
The answers were very different.
The biggest difference wasn’t presentation. It was how the platforms interpreted the transaction.
Claude surfaced terms that appeared in the documents but, in several cases, were historical or no longer controlling.
Notessa followed the transaction through subsequent documents and amendments to determine what the documents actually supported as current. And when the documents did not provide enough information to support an answer, Notessa didn’t fill in the gap.
That distinction matters.
In CRE, accuracy isn’t just finding a number somewhere in a document. It’s knowing which document controls, what changed later, and when the documents don’t support an answer at all.
Capability isn’t the same as productivity
Could a general-purpose AI eventually reach the same answer with more prompting, correction, and refinement?
Probably.
But that’s exactly the point.
Every additional prompt, correction, and instruction is more time spent teaching the AI how to do the job.
Multiply that across hundreds of assignments and dozens of employees, and the efficiency AI was supposed to create starts disappearing.
The goal shouldn’t be to make your people better at prompting AI.
It should be to make AI better at doing their work.
So maybe Noel Gallagher had it right. Don’t put your deal in the hands of just any AI and hope it figures out what matters. Put it in the hands of AI that already understands the work.
Put your work in the hands of a product built for the job.
Notessa. Built for commercial real estate. Designed for the way you work.