The people best placed to write your marketing copy are already emailing your support team.
We ran it on two WordPress plugins: WP RSS Aggregator, which brings content from other sites into WordPress, and Spotlight, which puts an Instagram feed on a website. Both have years of listing copy, landing pages and docs behind them, all written by the team, in the team’s words.
Buyers search, ask AI assistants and compare plugins in their own vocabulary. When that drifts away from the team’s, the right buyer can land on your page and fail to recognize their own problem in it.
The system we built
A customer language agent that reads support tickets and reports back in the customer’s words. It has five parts:
- Ticket pull. Six months of support conversations for both plugins, read-only.
- Customer text only. Replies, quoted threads, signatures, automated notices and vendor pitches stripped out, leaving only customers writing in their own words.
- Grouping in their words. Goals and situations clustered by how customers phrase them, each backed by verbatim quotes with names and sites removed.
- Their words against ours. How often customers use the terms the marketing uses, next to the terms they reach for instead.
- Headline candidates. Hooks built from real customer lines, each listed with the sentence it came from, so every claim traces back to a ticket.
What the tickets said
The vocabulary gap was the clearest finding. “Autoblogging” is in WP RSS Aggregator’s directory title, and not a single customer conversation used it. Customers “import”, “pull” and “fetch”, and they want a feed to “keep it updated”.
Spotlight followed the same pattern. “UGC” (user-generated content) and “social proof” appeared in zero customer messages. Customers want an Instagram feed on their website, and they talk about what shows, what’s connected and what updates.
The best lines described a situation, and a few needed almost no editing to become headlines:
- “We have a blog-site that generates all content for 2 other sites.” became “One blog that feeds all your other sites.”
- “Our users expect the website feed to match our Instagram profile.” became “Your website feed should match your Instagram profile.”
Some buyers already arrive through AI assistants. “ChatGPT recommended your plugin to me,” one wrote, and another had asked Google’s AI about a setting before getting in touch. Listing copy now has to answer the questions people put to an assistant as well as the ones they type into a search box.
What it produced
The report fed three pieces of work: directory listings for both plugins redrafted in customer phrasing, help articles titled the way customers ask the question, and a feature shortlist for the product team. It’s still at draft stage, so there’s no result to report yet. What has changed is where the words come from: every headline in the new drafts traces back to something a customer wrote.
Want the same for your business?
Your support inbox and sales calls already hold the words your buyers use. We build this agent on your own tickets or call transcripts and hand you the report.
