The 90-Day AI Visibility Roadmap I Run for Sub-DR-20 Sites

A 90-day AI visibility program for sub-DR-20 operators runs in four fixed phases: baseline and audit (days 1-14), entity graph plus five canonical pages (days 15-42), co-mention seeding (days 43-70), and measurement plus refresh (days 71-90). The phases sequence in this order because each depends on the previous one.

Every post in this cluster has argued that a sub-DR-20 brand can compete for AI citations by engineering entity coherence, publishing original data, and seeding cross-brand co-mention density. This post is the roadmap that packages those arguments into a ninety-day program an operator can actually run. It is the closing post of the cluster, and it is the entry point to doing the work. The pillar this roadmap implements is Answer Engine Optimization Explained.

The phases are fixed in order because each depends on the previous. An operator who ships canonical pages before fixing entity-graph coherence earns citations that resolve to someone else’s canonical. An operator who seeds co-mentions before the canonical pages exist wastes the seeds. The sequence is the program.

§1, Phase 01: Baseline and audit (days 01-14)

Read the current state. Run the schema-linter on every active page. Export the sameAs graph across LinkedIn, GitHub, Medium, Dev.to, YouTube, Twitter. Pull citation counts per engine on the current top five posts. Identify the named concepts the brand will own for the rest of the year. The output of phase 01 is a single document listing the drift, the missing surfaces, and the concepts.

§2, Phase 02: Entity graph plus five canonical pages (days 15-42)

Ship the Person schema with sameAs and knowsAbout. Ship the Organization and SoftwareApplication schemas with cross-referenced @ids. Publish llms.txt with an x-updated header. Publish /.well-known/llms.json. Write five canonical pages on the chosen named concepts. This is the heaviest phase, and it is the phase most operators skip by accident because the work looks editorial when the weight is actually structural.

§3, Phase 03: Co-mention seeding (days 43-70)

The phase where the moat compounds. Guest on one podcast that names the entity and the topic in the same sentence. Publish a GitHub repo with a README that includes Person schema. Comment on two Hacker News threads where the entity is technically relevant. Ship cross-brand citations in both directions. Cross-post to Medium and Dev.to with canonical URLs pointing back to chudi.dev. These seeds take four to eight weeks to surface as AI citations.

§4, Phase 04: Measurement and refresh (days 71-90)

Four readouts. Citation count per engine. Named-concept citation attribution. Entity coherence pass rate. Cross-brand co-mention density. Against each, the first refresh pass, topic drift republishes, freshness bumps, any sameAs drift corrected. By day ninety, the scoreboard is live and the next ninety-day block is planned.

§5, Why does cadence discipline beat volume?

Five canonical pages shipped on schedule across the ninety days outperform twenty pages shipped in a burst. AI crawlers apply a spam heuristic to velocity spikes. A steady cadence reads as an active editorial source; a burst reads as automated content generation, which many crawlers discount or delay. The schedule in phase 02 is the schedule, not an aspiration.

§6, What does “complete” mean at day ninety?

The roadmap is completed at day ninety, not finished. The scoreboard continues. The refresh cadence continues. The entity graph stays in drift-watch. What changes at day ninety is that the operator has a measurable program, not an intention. The delta between intention and program is the entire value of the ninety days.

§7, Why this is the cluster’s closing post

Every other post in the cluster argues for one piece of the work. This one sequences the pieces. An operator who reads only this post gets the program. An operator who reads the whole cluster understands the reasoning behind each phase. Both are valid entry points.

If you would rather decide before building, B2B SEO agency versus AI search specialist tells you which problem you have, and the AI Discovery Review measures it for you.

Bridge

citability.dev is the instrumentation for phase 04 and the early-warning system for the decay patterns from the previous post. The roadmap is the thinking. The scorer is the scoreboard. Sub-DR-20 brands that win this game ship both at once.