intelligence isn't knowledge.

plug the smartest model you can find into your Odoo and ask it a plain question. revenue this month. it answers fast, and with total confidence. and there is a good chance it is wrong.

not because the model is dumb. because it does not know your business.

“revenue this month” sounds simple. in Odoo it is not. the same table holds customer invoices, vendor bills, refunds, and drafts. amount_total is a field, not a definition of revenue. the model was never told any of that, so it sums the wrong rows and hands you a number that looks right.

ask for your top salespeople and it may list your top customers. ask what you can sell and it counts stock that is already reserved. every one of these is a clever system being confident about a thing nobody taught it.

that is the part i keep coming back to. we all wait for the next model to fix this. a bigger model does not fix it. it just reaches the wrong answer faster and argues for it better. intelligence was never the missing piece. specific, checked knowledge of your data is.

and that knowledge has a shape. which records count. which fields mislead. how two tables actually join. whether an action is safe to take. none of it lives in the schema, and none of it is something a model can reason its way to on its own. it has to be given, and it has to be true.

so that is what we built. the Context Layer for Odoo is that knowledge, verified against Odoo 19 and handed to the model before it answers. it is in early access now. you can point it at the plain revenue question and watch it refuse to guess, and ask you what you actually meant instead.

a model without it is a very well-read stranger walking into your business for the first time, every single time. brilliant. sure of itself. and wrong until someone tells it how the place really works.

intelligence gets you a fast answer. knowledge is what makes it the right one.