Workflows that earn their keep.
The real power of the connection is chaining tools — read a gap, write the fix, save the draft, all in one conversation. Here are the plays worth stealing.
A single read or a single write is useful. Chaining them is where the connection pays for itself: you can read where you fall short, draft the fix, and save it — without opening a tab. Every play below is one conversation. Reads cost nothing; the generate and run steps spend credits at the same price they do in the app.
Reading a gap and drafting the fix in the same chat is faster than any dashboard round-trip — you never leave the conversation to switch screens. Keep the human in the loop for the publish: content saved through MCP is always a draft, so you review it before it goes live.
From an AI-visibility gap to a WordPress draft
The end-to-end play: find where competitors get cited and you don't, read the recommended fix, write the article, and save it to WordPress as a draft you approve. Chain get_ai_visibility and get_ai_visibility_actions to read the gap and the diagnosis, get_search_ai_bridge to see which searches feed it, generate_blog_article to write, and publish_to_wordpress to save the draft.
- "Where do my competitors get cited but I don't, for my brand?"
- "What's the highest-leverage action from my latest AI-visibility check?"
- "Write an article that closes the biggest gap and save it to WordPress as a draft."
The article step spends credits (10 to 50, by length) and runs as a job — the assistant hands you a job id and polls until it finishes. The WordPress step never publishes a live post: it saves a draft, and you press publish yourself.
A Monday rival roundup
Start the week by summarising what your competitors shipped, then turn the sharpest item into a post. Read with list_recent_signals, get_competitor and get_daily_brief, then write with generate_social_post.
- "Summarise what my competitors shipped in the last seven days."
- "Which of those matters most to my customers, and why?"
- "Draft a LinkedIn post responding to the most important one, in my brand voice."
The reads are free. Only the last step spends credits — a social post is 2 credits, the same as in the app.
Close an SEO fix
Read the audit, pick the fix with the most upside, draft the content it needs, then re-check to confirm. Read with get_seo_audit (or run a fresh one with run_seo_audit), pull the affected page with get_content or get_blog_article, write with generate_blog_article, then re-run the audit.
- "What are the top three fixes from my latest SEO audit?"
- "Draft the missing content for the highest-impact one."
- "Re-run the SEO audit and tell me if that fix moved the grade."
The first audit on each brand is free. Re-checks cost credits by depth. So read the stored audit with get_seo_audit when you only need the current numbers, and re-run when you've actually changed something worth measuring.
Prep a client report
For agencies: pull one brand's headline numbers into a single summary you can paste into a deck. Read with get_brand, get_ai_visibility, get_seo_audit, get_traffic and list_recent_signals — all free.
- "Give me this month's headline numbers for <brand> — AI-visibility, SEO grade and search traffic."
- "Note anything that changed by more than its confidence range."
- "List the three most significant competitor moves this month for that brand."
When you quote a figure in a report, carry its ± range with it — an AI-visibility rate is a range, position is an average not a rank, and search traffic lags a few days. Stripping the caveats turns an honest number into a false one. See understanding the numbers.
Tips for chaining
- Name the brand in your ask. If you track more than one, saying which brand you mean keeps the assistant from picking the wrong one.
- Ask for a price before a big check.
quote_ai_visibility_checkis free and tells you the credit cost before you commit to the run. - Grant the right access first. The generate and run steps need a Content or Full token — a Read-only token can only read. See scopes and access levels.
- Reconnect if a result looks stale. Clients cache the tool list; toggling the connection off and on, or starting a new chat, makes it re-read the tools.