Content growth for every client site.
Each week, for each client: rising queries pulled from their Search Console, the SERP checked for an angle, a post written and scored against an SEO checklist, interlinked with the client's existing articles — and published to their WordPress only after you approve the draft. The agent remembers what it published and what ranked.
Already have an account? Set this up on a client store →
Query “how long does cold brew concentrate last” +38% WoW, 640 impressions, no matching page. Top 3 results are generic; draft answers in the first paragraph, links to 3 existing posts. SEO score 78.
A weekly post per client is 6–8 hours each. Nobody has that.
Research, write, optimise, interlink, publish, track — per article, per client. So posts go out when there is time, target queries someone guessed, and are never revisited to see whether they ranked.
Hours per article
Even a seasoned content team struggles to ship weekly for one site. For ten client sites it becomes quarterly, then nothing.
Wrong topics
Without the client's Search Console data you're guessing at keywords. Most posts never rank because they target the wrong query — or one the client already ranks for.
No learning
Published and forgotten. Nobody checks which posts ranked and why, so every article starts from zero context and the client's blog never builds topical authority.
One content policy per client site. Publishing asks.
The schedule is the trigger. Research, drafting and scoring run inside the agent step; the draft lands in your inbox with the target query and the score. Nothing is published on a client's site without a sign-off — unless you've allowed autopilot for that client.
- Scoped to one client — its Search Console, its WordPress, its business profile and tone.
- If a rising query matches an existing post, the agent proposes an update to that post instead of competing with it.
- Every draft is scored — word count, keyword density, readability, headings, internal links, alt text — and revised until it clears 70 before it reaches you.
- After publishing, 2–3 existing posts get links back to the new one. Each is its own approved write.
Schedule → policy → agent → ledger.
Monday, per client
The policy fires on its schedule. On the first run the agent reads the client's site and saves a profile — product, audience, tone, conversion paths — and references it in every article after.
Topics from real data
Rising queries, top performers and low-CTR opportunities from the client's Search Console, mapped against their existing posts into a topic cluster. Thresholds you set decide what's worth a post.
Research, write, score, propose
Searches the target query, reads the top 3–5 results for format and weaknesses, picks an angle, writes the post — direct answer first, keyword-targeted headings, FAQ, natural product CTAs — scores it, and proposes publish with the evidence attached.
Remembers and reports
What was published, which assumptions were made, which posts gained position. Next run starts by checking whether last week's assumptions held. The client's monthly report shows posts, positions and who approved them.
Typed content actions — not a “generate article” button.
You approve “publish post 2291 on harborandco.com”, never “let the AI run the blog”. Reads of Search Console and existing posts are free. Writes to the client's WordPress ask by default.
- toolPull rising queries & low-CTR pages from Search Console
- toolMap existing posts into a topic cluster; prevent cannibalisation
- toolResearch the SERP for the target query
- toolWrite and score the draft; revise until it clears the bar
- toolPublish or update a post on the client's WordPress
- toolAdd internal links from existing posts; notify your #content channel
Put every client's content on a weekly policy.
Connect Search Console and WordPress for the first client, attach the content policy, approve the first draft. Free for one site; agency plans from the second.
Free for one store · 14-day trial on paid plans · no card · bring your own AI key