Per-template JSON-LD
I write the schema generation code for every priority template, ship it through your CI, and pair with your engineers so they can extend it after I'm out.
What it actually is
For every template in scope (typically 10 to 50), I write the JSON-LD generation code in your stack. Server-rendered. Typed where your codebase is typed. Tested where your codebase has tests. The schema gets emitted in the initial HTML response, not injected client-side, so AI engine fetchers see it without executing JavaScript.
Agencies that hand off schema usually deliver a JSON file or a CMS configuration. I deliver code that lives in your repository, follows your conventions, and gets reviewed by your engineers. By Sprint end your team can add a new template's schema without calling me. That's the bar.
Deliverables
- Production-merged PRs for every template in scope
- Reusable schema-generation utility shared across templates
- Type definitions or interfaces for every Schema.org type used
- Unit tests for the generation logic
- Per-template before/after validator output
What breaks without it
Schema implementation that isn't in the initial server response is invisible to most AI engine fetchers. Schema implementation that lives only in a CMS GUI is impossible for engineering to test and review. Both are common, and both produce the same outcome: schema that scores fine in a one-off check and decays the moment a developer ships an unrelated change.
The recurring failure I see: a previous vendor implemented schema as a tag-manager template. It worked when they tested it. Six months later a separate marketing change broke the tag, nobody noticed, AI citations dropped, and there's no diff in source control to point to. Implementation in the codebase makes that failure mode impossible.
How it fits the Sprint
Per-template implementation is where the architecture meets production. The validator suite gates these PRs. The pair sessions teach your team to extend them. The pre/post measurement compares against the AI visibility baseline. Every other Sprint sub-service exists in service of getting these templates right.
The full Sprint breakdown
Stop pouring budget into a broken foundation.
If your SEO retainer hasn’t compounded, your AI citations have stalled, or your last technical audit ended in a deck nobody read, that’s not a content problem. It’s an engineering problem. The same engineer who diagnoses ships the fix.