The Content Plan: how the 15-item plan is built and explained

Updated 22 October 2026 (Agent 11.1: change feed folded in: #3593, #3594 live; country spelling, plain titles and no patient stories, #3924). Previously updated 7 October 2026 (Agent 11.1: catch-up from every lane after their state-doc updates; scoring rebuild of 21 Sep, no score-trend claims, no citation claims for schema or credentials, Sites, content plan, support). Plain-English briefing on My Plan: what the customer sees, how the plan decides what goes on it and in what order, and why no two clinics get the same plan. Written 24 July 2026. Rewritten 2 October 2026 (Agent 11.1) from platforms/ccs/state/content-plan.md, "HOW THE CONTENT PLAN WORKS" (Agent 1.1 with Tor, 2 October 2026). That state doc is the source of truth; where this briefing and it disagree, the state doc wins.

What the customer sees

Every clinic gets My Plan: 15 items, numbered 1 to 15 in order, each one a concrete piece of content that will improve the clinic's AI visibility. Each item is either a new page or a rewrite of a page already on the clinic's site, at the same web address, never an extra page. Page types are condition pages, service pages, treatment pages and blog posts; a rewrite keeps the type of the page it replaces. Profile cards appear at the top for any clinician below 60% profile completeness, because that is the level needed to author content.

The items sit under five headings, in this order:

1. Your Practice Priorities: up to three topics the clinic wants to be known for, all counting equally.
2. Your top services: the first five on the clinic's ranked services list, in its order.
3. Your top conditions: the first five on the ranked conditions list, in order. Services come before conditions.
4. Your clinicians' specialisms: anything ranked sixth or lower, and expertise the clinicians hold that the clinic has not ranked.
5. Building out your topics: the remaining work that rounds out each topic area.

Every item has a Write it for me button that starts the writing. The plan is not advice; it is a queue of work the system does.

How the plan decides what goes in, and in what order

Coverage first. Every Practice Priority, every top-five service and every top-five condition gets a place before anything else is added. Usually that place is the topic's main page. If something stops it (we published that page for the clinic recently, the clinic marked it "Not relevant", or nothing eligible exists), the plan records the real reason rather than guessing.

Then the rest, in heading order. No topic can take more than three places in the plan, so one subject can never take it over.

The clinicians decide the order within each heading. A starred clinician counts most and lifts everything that person treats. Next come the items a clinician has marked as one of their key specialisms, then anything else they list. The more of the clinic's clinicians who share a specialism, the stronger it counts.

Market evidence orders the work within a topic, never lifting it above a higher heading: nearby competitors that AI tools name or cite for the topic when they don't name the clinic (the strongest signal), search volume, and the questions patients actually ask. Within a topic, the main page comes before its supporting articles.

The clinic's own existing pages can be its biggest quick wins. A page already on page one or two of Google for a topic the clinic cares about, a page slipping down Google, or a page on a topic where AI tools recommend a nearby competitor instead, gets a strong lift within its heading. Rewriting a page that is already close is the fastest route to being cited by AI. Pages already working are protected: a page in Google's top three, or a page AI tools already cite for the clinic, is never offered for a full rewrite. The homepage is never offered as a rewrite.

Opportunities. At most two per plan, badged "Opportunity": a topic outside the clinic's lists, offered only when AI tools recommend nearby competitors for it, don't recommend the clinic, and one of the clinic's clinicians already treats it.

Pieces we have written leave the plan once published. They come back for a refresh no sooner than 12 weeks later, or sooner if the clinical evidence behind them changes, such as a new guideline. Refreshes of our pieces come after new work, weakest page first compared with the rest of the clinic's own site. The clinic's own pages (never written through us) are different: they compete with new work from day one.

Started work is never removed. An item the clinic has started keeps its place and progress.

Focus mode builds the plan from the clinic's Practice Priorities only. Profile cards, changed-evidence refreshes and started work stay.

Every item explains itself

Each card's "Why this is in your plan" explanation is built from a saved record of exactly why the item is there. It names the evidence: the AI tool and the question asked, the competitors it recommended instead of the clinic, the Google position, the priority or ranking it serves, and the clinicians by name. For example: When we asked ChatGPT "who treats running knee pain near me", it recommended two nearby clinics, but not you. Nothing is generic and nothing is "the algorithm said so". The full record is kept for every plan, so support can always show why one item placed above another.

Why no two clinics get the same plan

• The clinic's own priorities and rankings set the order. The plan can only ever be a sharper version of the clinic's own strategy.
• The clinic's people decide what can be written. Topics are anchored to clinicians' declared expertise; the plan cannot recommend content nobody on the team is qualified to author.
• The clinic's existing site shapes everything: what already exists, what scores well, what ranks, what AI already cites. (Legal and boilerplate pages are excluded from the customer-facing page lists; they are not content.)
• The clinic's own area and competitors shape the evidence: the AI questions are asked about the clinic's area, so the competitors named are the clinic's real local competitors.
• The customer steers it. Any suggestion can be marked "Not relevant", with a reason and an optional note. The suggestion is removed, and similar suggestions with the same reason are held back for 90 days. The rejection record itself is never deleted.
• The clinic's country changes the content itself (see the research briefing).

Keeping the plan current

• Automatic rebuild after priority or top-five changes (live 3 Oct, #3533/#3535). Changing Practice Priorities or the top five services or conditions rebuilds the plan about an hour after the last change: at most once a day, no extra charge, no new research (it re-orders using what we already know), started work kept. Don't encourage clinics to keep reshuffling lists to get new suggestions (Tor). Other changes still wait for the Refresh Content Plan button.
• Primary (live, #3521/#3522/#3528). Each list (niche, condition, technique) can have one Primary; it counts as much as a starred clinician, locks for 90 days once set, and starts full research. Practice Priorities stay active after their 90-day lock ends; the lock only stops editing.
• Cards (deployed 5 Oct, #3542/#3544). At most one search figure per card: "Around [N] people in [Country] search for "[keyword]" each month." Cards say "It relates to your Practice Priority, [x]." only when it genuinely does, and "[Name] treats this…" for a named clinician. One main page per topic, and no new main page for a topic already covered under a slightly different title (#3555).
• Rewrites and failures (live 6 Oct). A rewrite of an existing page writes about the page's real subject, starts from the current page and uses the clinician's answers (#3551). An incomplete article shows "Generation failed" with a retry, never raw code, and nothing broken overwrites a good version; automatic quality repairs keep the best version and are not charged; credits are taken once per piece when requested (#3556/#3559).
• Started work keeps its place (live, #3593). A rebuild keeps every piece the clinic has started in exactly the same place.
• Refresh says what happened (live, #3594). While it runs: "Rebuilding your plan…". If the rebuild can't go through, a message says the plan was not changed, and the previous plan stays on screen with "Last saved plan: [date]".
• Card titles and subjects (live 20 Oct, #3924 and the 19 to 22 Oct prompt changes). Titles use the clinic's country spelling ("Program" in Australia) and plain words ("Neck-Related Headaches", not "Cervicogenic"). The plan never suggests patient stories, case studies, testimonials or success stories, titles that promise a result or a speed, or titles that claim one treatment beats another. It doesn't suggest a subject the clinic has already published. Existing plans pick this up at their next rebuild.
• Still NOT built: "Not relevant" shaping the plan, own-topic requests, "include / leave out" boxes and outline approval.
• The plan changes when it is rebuilt, not the instant a setting changes. A significant change (priorities, ranked lists, clinicians, profile edits, the clinic profile) marks the plan as due a refresh, and the Refresh Content Plan button rebuilds it. Progress shows while it works; if a rebuild fails, an honest "regeneration failed, try again" notice appears with a retry button, and the plan shown is clearly labelled as the last successful version.
• A new assessment updates the evidence but does not rebuild the plan on its own. The plan uses the new evidence the next time it is refreshed. Say "every time your plan is refreshed", never "every month".
• The plan never dries up. When five or fewer actionable items remain, it rebuilds itself in the background (at most once a day) and refills to 15. The customer may briefly see "Adding fresh suggestions…".
• Progress is visible. Cards show live stages (Not started, Waiting for insight, Research in progress, Written, Peer Review, Published). Published items leave the plan; published work lives in the Content Studio.

Author input survives plan changes. Answers a clinician submits on a plan item are saved against that item and survive refreshes and rebuilds. The card shows a green "Author insight received · Ready to write" badge with a read-only "View submitted responses" view, and writing starts from the saved answers via Write it for me. The "Author input received" email shows the real article title and links straight to the item, including through magic-link sign-in (#2845/#2846, live 30 Aug 2026). A customer who lands on an internal Co-Kinetic screen is redirected to their own My Plan (#2847, live 31 Aug 2026).

Common questions this briefing answers

• "Is this just a list of blog ideas?" No. Every item comes from the clinic's own priorities, rankings and clinicians, ordered by measured evidence, explained on its card, and written by the system at the press of a button.
• "Why isn't my number one service at the top?" Practice Priorities come first, then services. Every top-five service is guaranteed a place unless there is a stated reason, and the card says which.
• "Why is my existing page in my plan when I asked for new articles?" The clinic's own pages compete with new work, and some are its quickest wins. Rewriting replaces the text at the same address; it never adds a page.
• "Will you rewrite a page that's doing well?" No. Pages in Google's top three, and pages AI tools already cite, are left alone.
• "We published an article; why isn't it being refreshed?" Pieces we write leave the plan once published and come back for a refresh no sooner than 12 weeks later, unless the clinical evidence behind them changes.
• "Where do the competitor names come from?" We ask AI tools (ChatGPT, Google AI Overviews, Perplexity, Claude) the questions patients ask about that topic in the clinic's area, and record which clinics they recommend or quote. A card only shows answers about its own topic.
• "Can I add my own topic?" Not yet; it is coming soon. Never give a date.
• "I re-ran my assessment; why didn't my plan change?" A new assessment updates the evidence; the plan uses it the next time it is refreshed.
• "How do I get location-specific service pages?" They are a different page type from anything in the plan today and are coming soon. Never give a date.