The property description got rewritten by the agent and I think it cost us a week

The pattern you're describing is real, and the accountability gap is structural. Most listing content vendors deliver a final file, at which point their involvement ends. The agent owns the listing, so the agent controls what goes live. There is no standard mechanism in the industry for a vendor to track post-delivery edits, and most vendors have no contractual right to one either.

A small number of shops are starting to build what you might call a change-log layer into their workflow, usually by hosting the copy in a shared document with version history visible to both parties, or by delivering through a platform that timestamps every edit. A few are pairing delivery with a short briefing where they walk the agent through why specific choices were made, on the theory that an agent who understands the logic is less likely to rewrite on instinct. Neither approach is common, and I have not seen data on whether either actually reduces agent rewrites. My honest uncertainty here is about adoption: I would want to ask vendors directly whether they offer this and how many agents actually use it, rather than assume it is widespread.

The post-settlement environment you are watching from the outside does raise the stakes for agents to demonstrate value, which in theory makes professional differentiated copy worth protecting. Whether that pressure translates into agents being more willing to leave vendor copy intact is something I genuinely do not know.

Your version-history instinct is the right frame. The question is who controls the repo.

Are you thinking about this as a buyer of listing content services for your portfolio, or more as someone who might build or offer that accountability layer as a product?

The version control problem you're describing is real and mostly unsolved on the vendor side. The dynamic you saw, specific lead replaced with "welcome to this charming," is the most common single failure point in listing copy, and it almost always happens after delivery rather than before.

Most listing content shops deliver a finished asset and lose custody of it at that point. Some include a brief on the brief, a document that explains the structural choices so the agent understands why the hook leads, why the differentiator is in sentence one, what gets lost if that order changes. Whether agents read it is a different question. A few shops have moved toward a light review step where the agent submits edits for a quick pass before the revised version goes live, but that requires the agent to see the vendor as a collaborator rather than a fulfillment service, and that relationship usually has to be sold, not assumed.

The risk you didn't name is attribution. If the copy performs badly in the version the agent published, the vendor's portfolio reflects a result they didn't produce. Sophisticated shops are starting to document the delivered version separately from the published version, at least for their own records. Whether they have contractual language that addresses what happens to the copy post-delivery is something you'd want to see in the service agreement, not just in conversation. That's a point where an attorney should review any terms if you're building a vendor relationship with that kind of IP or performance clause.

Your traffic data insight is the sharpest thing in this thread. The first sentence is a filter, and burial is a structural failure, not a voice preference.

Are the agents you're seeing this with using listing copy vendors on retainer, or as one-off orders per listing? That relationship structure probably determines how much leverage a vendor has to build any kind of gate.