Now draw lines between all the circles. Those are your lines of communication without a PM. This is a sad diagram, no one knows the whole story.
Now repeat the exercise but draw an additional circle in the middle of the page.
This time, the lines on the left and right connect to the middle circle. This is a happy diagram, the developers aren’t pulled in a dozen different directions and one person has the whole picture.
That is the role of a PM.
In the example, every one of those lines consumes resources: time, context switching, misunderstanding, synchronization with the rest of the team.
A quality PM funnels all those lines through themselves. Their value is proportional to their ability to (a) bidirectionally extract, transform, and deliver useful information to both sides (developers and end users), (b) offload processing tasks that neither side is good at (merging similar requests, clarifying ambiguous requests, tracking down answers / requests for external teams), & (c) keeping a 10,000 view and coordinating tactical requests to align with strategic vision.
I like the CDC 6600's peripheral processors ( https://en.wikipedia.org/wiki/CDC_6600#Peripheral_Processors... ) as a computer hardware metaphor.
The PM is the hub of communication between all 4.
Although on a day-to-day basis, it's the dev team and business/sales that make up the large majority of interaction.
Check-ins with management are occasional, and conversations with customers are just kind of an ongoing informative thing.
There’s a lot of assumption built into the framing of your question. Why do you think a PM is a middleman and only creating meetings? Is this based on some experience? Why is your premise that the business team owns the roadmap? Bad PMs do exist, just like bad managers, and just like bad engineers. If that’s where your question is coming from, I’m sorry, but please don’t assume that a bad experience is universal; it’s not.
In a business, there usually needs to be a healthy tension between what sales and marketing ask for, what the executive leadership see in the future, what customers need (which is not always what sales and marketing are asking for), what is on the engineering backlog, how difficult features are to implement (both how long, and also how disruptive they are), and anything the product needs that is not on one of those previous lists. When one team or one point of view dominates, bad things tend to happen. If sales and marketing gets to dictate what happens and engineering is relegated to implementing whatever they say, the roadmap can be too short-term and full of churn, it will lack focus and forget about existing customers. If engineering has complete control over the roadmap, it sometimes drifts away from what customers need and toward what engineers like to do, with solutions that are neat but overly complicated or using new technology for the sake of learning. You need a little of each, as well as some pushback from every direction to keep everyone happy. A PM has their work cut out for them.