Opinions should be structured based on a series of cause-effect relationships. This applies to your opinions and others.
If your opinion is “AI is exciting, we should use AI in our product”, this is an unstructured opinion. It’s weak.
The conclusion of your opinion maybe correct; “we should use AI in our product”, but not because it’s exciting. But instead because of a structured cause-effect relationship.
Humans are expensive, companies will invest in technology that can do parts of a human’s job, these parts of a human’s job are semi-structured, AI works well on this kind of data, we should use AI in our product.
The beauty of a logical cause-effect chain in an opinion is that we can interrogate the reasoning behind the opinion. We can know if the reasoning, and by extension the opinion, is sound. A single assumption in the cause-effect chain being wrong shows how the conclusion is wrong. If the whole chain of assumptions holds, the conclusion should hold.