This does happen, and it's usually the result of a communication / execution dysfunction. To start with, no salesperson wakes up and says let me go around and undermine the PM - they usually have trying to follow a process that doesn't work and eventually take matters into their own hands becuase they are accountable to the business for real revenue. Some of the things I'd diagnose if faced with this situation:
I think the PM role can be very narrow when there are many fully staffed cross-functional partners with expertise across engineering / UX / testing / data / marketing, which leaves the PM as a dot connector and customer proxy (i.e. large company). And when many of those partner functions don't exist (i.e. small company), the PM can become a jack-of-all-trades.
Neither is better per se, they just offer different learning opportunitie.
There are obviously a bunch of goal setting / tracking frameworks that are designed to do this (for example OKRs) but I tend to start with where and why any misalignment exists - usually there are 4 levers, and they all come back to strategy:
My # goal is to provide leverage around the strategy:
I try and start with a combined top-down / bottom-up exercise.
Top-down: what are the top 1-3 business objectives that we can move the needle on with product work? Let's prioritize those.
Bottom-up: what is the bare minimum keep-the-lights-on staffing level required to run our service / business?
If there is any surplus from this exercise, you take on more of the backlog and can do more ambitious thing. If there is a deficit, time to deprioritize certain investment areas until the team is staffed to a healthy level.
I would start with minor tweaks to the existing process vs trying to completely re-do things - most changes that stick only tweak 1 variable at a time vs trying to change the whole system. So I would certainly not come in and try to redefine how ideation, scoping, planning, commits, shipping, reviews, etc work on Day 1. But if the team already ideates then I'd try to introduce customer conversations / user research as another avenue to generate ideas, then connect the dots to show that ideas that were grounded in external learning (vs internal opinions) had higher value over time.
The biggest challenge comes from cross-functional partners (engineering, sales, marketing) who may not have prior experience partnering with PMs and aren't bought in to introducing the function. The real value of a PM is to have a dedicated resource thinking about what the highest-leverage activities fro the team are in service of the broader business goals and user problems, but in many orgs that role has been played piecemeal by various folks and solidifying into a formal function can feel premature / overkill / threatening.
Additionally, when introducing PMs into a company that hasn't had the role before, you also have an external challenge of creating enough of a brand to attract talent to hire.
Ideally, you have "vertical" product teams whose charter is shipping end-user-facing features around a particular problem / persona. You may also have "horizontal" product teams that own certain shared services / infrastructure / components that multiple vertical teams leverage. These teams will usually have some combination of a engineering / product / design (EPD) functions, and may also include user research, data science, and/or quality assurance.