Content Engine & Content Studio Requirements

Content generation, blog optimisation, brand guidelines, content buckets and Content Studio roadmap requirements. Consolidated for planning and reference.

Milestone: 1 · Date: 22 August 2026 · Status: ⬜ Planned Back to milestones · roadmap map

Capability flow

brand guidelines
        |
content configuration
        |
dynamic inputs
        |
content generation
        |
AI / quality evaluation
        |
SEO optimisation
        |
review
        |
final content output

1. Blog content generation & optimisation

1.1 Blog tonality

Tonality guidelines and filters apply before generation. The workflow incorporates the selected tonality at the generation stage, not as a post-processing pass.

1.2 Keyword strategy

Tier	Count	Selection
Primary	1-2	Dynamic, by search volume
Secondary	3-5	Dynamic, by search volume

The system determines the primary and secondary keywords as part of the generation/optimisation workflow. The keyword counts are not fixed — an admin can set them dynamically; 1-2 primary and 3-5 secondary are the defaults.

1.3 Blog title in thumbnail

The blog title is incorporated into the blog thumbnail. The generation prompt is updated to support it.

1.4 Dynamic AI detection score

A dynamic AI detection score for generated blog content, produced at generation time and available as part of the content evaluation workflow.

1.5 AI phrase identification

AI-generated or AI-like phrases are identified within the content. Anything above 15% is highlighted for review. This is a review mechanism — it flags, it does not silently rewrite the generated content.

1.6 Inline sources

Inline sources and citations are removed from the generated blog body where they are currently inserted directly into the article.

1.7 Internal blog interlinking

Generated blogs support internal interlinking with relevant existing blogs. The system identifies suitable internal references and inserts links where appropriate.

1.8 Scoring documentation

A new scoring document is implemented, and becomes the reference for the revised content evaluation/scoring framework.

2. Parent brand guidelines & content scale

Parent brand guideline. A parent-level guideline is the overarching reference for generation. Individual product and content requirements operate inside the boundaries it defines.

Product-level guidelines.

Dimension	Scale
Products	~54
Social pages per product	3-6
Guidelines	May differ per product

Guideline-based generation. Content is generated and structured according to the guidelines that apply to each product — the correct guideline at the relevant product/content level, not one generic format for everything.

3. Content buckets

The Content Studio supports multiple content buckets and formats, including but not limited to:

Infographics
Educational
Other defined content formats as required

Each bucket carries its own generation requirements and output structure.

3.1 Chart-based content

Prompts define what the chart needs to communicate and how it is generated.
The data required to generate the chart is provided/defined separately.
The system distinguishes the chart instruction from the chart data.

4. Content Studio as a SaaS platform

Product direction. The Content Studio is a SaaS product, not only an internal content-generation interface. Architecture and roadmap account for scalability and reusable generation workflows.

Carousel content. Slide count is configurable per content requirement.

Content inputs / references. Generation supports configurable parameters where relevant to the selected content type:

Tonality
Reference images
News
Videos

5. Key product direction

A configurable, guideline-driven Content Studio that generates multiple content formats while incorporating brand rules, product-specific requirements, dynamic SEO inputs, content-quality evaluation and reusable content references.

6. Platform & operations

Added 25 August 2026, assigned to Team 1 alongside the blog targets (§1.1-1.6):

Unit economics. Figure out the cost model — what a tenant, a generated piece and the AI/scraping spend behind it actually cost and return.
Platform monitoring dashboard. A dashboard surfacing bugs and platform health for the operating team.
Global/National Discovery filters. Discovery filters by scope — global vs. national — extending the existing statewise (geoState) filtering.

Added 27 August 2026:

Guaranteed blog FAQs. Every post ends with a fixed number of reader FAQs. The count was requested in the prompt but never enforced, so a model that returned fewer shipped a short post behind a warning; the shortfall is now a repairable quality issue. The count moves out of code into Settings → Blog generation, alongside the AI-score and keyword-tier knobs.
Published-blog link hygiene and in-body sections. Outbound LLM-chosen links carried no nofollow, passing equity to unreviewed domains; internal and same-page links are deliberately left to follow. FAQs and Sources were rendered only at delivery, so the reviewer could not see or edit them in place — both are now baked into the body after every quality gate.
Manual blog creation. A no-Discovery path: a human supplies a topic, keywords and reference URLs on its own compose page, and the existing queue and generator handle the rest. Reference extraction drops low-quality articles rather than generating from thin source material.

Targets

Targets tagged team:1 land in Team 1's charter column; untagged targets belong to Team 2. Within a team, dev:N splits the card by developer.