Agreed, and I think following flowchart-type logic is within today's AI capabilities. This thread is full of people getting inaccurate responses from humans. I think when it comes to accuracy, a well-trained LLM likely beats the status quo of high-churn low-paid employees following a rote diagram.

Of course there should always be a way to reach a human, a senior agent with actual knowledge that can be applied in subjective ways to solve more complex problems.