The AINEO Visibility Production Line: One Engine for Text, Video and Social

What the line does, from topic to published page

The line performs six operations in a fixed order: define the topic and who searches for it, split that topic across intent axes, assign each axis to an identity. Then write to the standard skeleton, check against the gate, and publish on a schedule. Nothing skips a step. A piece that fails the gate returns to its writer rather than to the queue.

The topic sheet comes first

Before a word is written, one page records the main query, who is searching, whether any existing page already ranks for that query, and the language. The system treats a missing answer here as a blocker. Topics without a defined audience do not enter production.

The output is a set, not an article

One topic produces eighteen pieces in each language, mapped so that no two pieces compete for the same search. The architecture those pieces land in is described under website visibility.

One topic, nine intent axes

The 1K9N standard splits a topic by search intent before it splits by format. The 1K is who: the person, the audience, the community searching. The nine N axes are what, how, why, how much, where, from where, when, what happened and what will happen. Each axis becomes its own page with its own owner.

Why the split happens before writing

Writers given a topic produce overlapping articles. Writers given an axis produce distinct ones. The how-much axis carries price and cost intent, the from-where axis carries data and source intent, and neither gets published without a named source.

Balance rules inside a set

Every axis appears at least once in a set of eighteen and none appears more than three times. The how axis is the single exception at three, because process questions genuinely divide. The search-intent reasoning behind the AI-facing axes is covered in what is ai seo.

Twelve identities, and where the human sits

Production is distributed across twelve declared identities: four real signatures, two brand voices and six AI-assisted desks. The four signatures are real people who write and approve. The two brand voices are Adapte Dijital and AINEO. The six desks are Research, Guide, Comparison, Local, News and Story, and each desk is paired with a named human editor.

What each identity type is for

Pillar and hub pages carry a real signature, because they make commitments. Support pages go to the identity whose craft fits the axis: process pages to the systems writer, case pages to the storyteller, data pages to the Research desk. The twelve lines of delivery those pages support are mapped under visibility services.

AI sits beside the writer, not in place of one

The desks are not pretend people. Each one is a craft with a real editor attached, and the editor’s name appears on the piece. AI assists with research, structure and drafting. A real signature does the final read, and that read is the point at which a piece becomes publishable.

The quota does not stretch

If a topic does not fit a desk, that row moves to the nearest real signature instead of inventing a new one. The daily quota stays fixed and the overflow moves to the next day. Google’s own spam policies for web search name scaled content abuse as generating many pages without adding value. A fixed quota and a hard gate exist to prevent exactly that.

The FORM-97 skeleton and the graphics standard

FORM-97 is the written standard every piece is built on. It fixes the title frame, the opening answer, the H2 shape, the length band by page type, the card architecture and the image rules. It also bans specific patterns, including emoji, bold inflation, repeated skeletons across a set and length padding for its own sake.

What the skeleton fixes

The first two paragraphs answer the question completely, and the first hundred and twenty words have to stand alone as a quotable answer. Every H2 is a question or a clear claim, followed by a direct answer paragraph before any expansion. The H2 sequence changes from piece to piece inside a set.

A cover and three graphics per piece

Every article ships with one cover image and three in-article graphics, produced with the article rather than afterward. The graphics are built from the article’s own lists, steps, tables and figures, so no graphic carries information the page does not. Where AI-assisted video fits the same batch is compared in ai video vs production.

Bekçi: a gate that can refuse a draft

Bekçi is the quality gate, and it checks three things before anything publishes. First, whether another page already owns the target query. Second, how similar the draft is to the other pieces in its own set. Third, whether it scores at least 80 out of 100 on the standard’s checklist, since a draft below that floor goes back to its identity.

Query ownership comes first

One query has exactly one owner page. Bekçi runs this check before writing starts and again before publishing, because titles drift during drafting. A collision found at the gate is cheap, while the same collision found six months later costs a merge and a redirect.

In-set similarity, measured not guessed

Pieces inside one set get compared against each other for overlapping phrasing and overlapping angles. The target is low overlap across the whole set, which is why eighteen pieces on one topic do not read as eighteen versions of one piece.

The threshold is a floor, not a grade

Eighty out of a hundred is the publish floor, not a target to celebrate. Pieces often clear it and still get rejected by the editor for a reason the checklist cannot see, such as a weak example or a claim without a source. The gate removes the mechanical failures so the human can spend attention on judgment.

What runs today and what is being built

Four parts of the line are in daily use: the FORM-97 skeleton, the 1K9N axis distribution, the twelve-identity assignment and the cover plus three graphics standard. Timed publishing across the day runs today as well, scheduled piece by piece. Bekçi runs today as a checklist applied by hand. Two distribution engines are still in build.

AINEO Socialsys, in build

Socialsys takes a published article and produces a short share cut from it: a sixty to ninety character headline, a description, tags and an image pulled from the article itself. It then distributes those cuts in a set order across the platforms that offer an API. The social calendar it is designed to feed is described in how to manage social media.

AINEO NEXUS, in build

NEXUS reads the Search Console export for a target page, scans the other pieces on the site for those queries, lists link candidates, then applies the approved ones. It is the engine behind the link map rather than a replacement for deciding what should link to what.

A worked example

Say a solar installer in Denizli becomes the topic. The how-much axis goes to the systems writer as a cost page, and the where axis goes to the Local desk. The from-where axis goes to the Research desk with sourced figures, and the pillar carries a real signature. Bekçi checks that the existing page for the main query is either the pillar or gets narrowed, then the set publishes across one day and returns in three months for its update pass.

Frequently Asked Questions

Short Glossary

1K9N

1K9N is the name for the distribution standard that splits one topic across who is searching plus nine intent axes.

FORM-97

FORM-97 is the written content standard used for structure, title frame, length bands, card types and image rules.

Bekçi

Bekçi is the term for the quality gate that checks query ownership, in-set similarity and the publish score before anything goes live.

Next Step

Pick one topic you already compete for and run the first two stages on paper: write the topic sheet, then split it across the nine axes and name the owner page for each. That exercise alone usually surfaces two pages of yours competing for one query. The delivery standard that sits around this line is set out under visibility service deliverables.

Updated: October 2026