The three usual failures
The disguised demo. A speaker with something to sell gives a talk shaped like an argument that resolves into their product. The tell is that every example points the same way.
The wrong altitude. Too technical and half the room stops listening in the first ten minutes; too shallow and the engineers write it off, which they will tell everyone else at lunch. Getting this right is a question about your audience, not about the speaker.
The talk with a short shelf life. A list of what is impressive this quarter is out of date by the time your video goes up. An argument about how to think outlasts its own examples.
What to listen for instead
Whether the speaker does this work or reports on it. Both can be good, but they are different talks, and a room full of practitioners can tell within two minutes.
Whether they will say what they do not know. A speaker who caveats is a speaker whose confident claims are worth something.
Whether the examples are theirs. Second-hand case studies survive contact with an audience right up until somebody asks a follow-up question.
Four questions to ask before you book
What will the room be able to do on Monday that they could not do on Friday? A speaker who has thought about the talk has an answer.
Which parts of this will be out of date in a year, and which will not?
What do you need from us to make this specific to our audience?
What will you not talk about? The answer tells you where the honesty is.