<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Insights — NOKARIS</title>
    <link>https://nokaris.ca/en/insights</link>
    <atom:link href="https://nokaris.ca/en/rss.xml" rel="self" type="application/rss+xml" />
    <description>Essays and takes on headless commerce, applied AI in production, web performance and e-commerce engineering.</description>
    <language>en-CA</language>
    <lastBuildDate>Thu, 20 Aug 2026 00:00:00 GMT</lastBuildDate>
    <item>
      <title>In a headless rebuild, the bottleneck isn&apos;t the front end — it&apos;s the content model</title>
      <link>https://nokaris.ca/en/insights/content-model-is-the-bottleneck</link>
      <guid isPermaLink="true">https://nokaris.ca/en/insights/content-model-is-the-bottleneck</guid>
      <description>You pick a headless CMS off a feature matrix, then discover the real work was elsewhere: deciding what a page is, and how marketing publishes without going through a developer. Content modelling is the most underestimated part of a rebuild.</description>
      <pubDate>Thu, 20 Aug 2026 00:00:00 GMT</pubDate>
      <category>headless-commerce</category>
    </item>
    <item>
      <title>An AI project doesn&apos;t die from a bad model, it dies from a bad scope</title>
      <link>https://nokaris.ca/en/insights/ai-projects-die-of-bad-scope</link>
      <guid isPermaLink="true">https://nokaris.ca/en/insights/ai-projects-die-of-bad-scope</guid>
      <description>Enterprise AI projects almost never fail on model quality. They fail because the scope had no owner, no measurement, and no acceptable failure threshold. Here are the four tests I put a use case through before writing the first line.</description>
      <pubDate>Thu, 13 Aug 2026 00:00:00 GMT</pubDate>
      <category>ai-applied</category>
    </item>
    <item>
      <title>Why most Hydrogen migrations take 6 months instead of 90 days</title>
      <link>https://nokaris.ca/en/insights/why-hydrogen-migrations-slip</link>
      <guid isPermaLink="true">https://nokaris.ca/en/insights/why-hydrogen-migrations-slip</guid>
      <description>Ninety days is achievable. What blows it up is almost never the storefront build: it&apos;s the Shopify app ecosystem you discover halfway through, the redesign grafted on along the way, and the content model nobody audited before quoting.</description>
      <pubDate>Thu, 06 Aug 2026 00:00:00 GMT</pubDate>
      <category>headless-commerce</category>
    </item>
    <item>
      <title>NestJS and Angular: the foundation for business applications that last</title>
      <link>https://nokaris.ca/en/insights/nestjs-angular-for-business-apps</link>
      <guid isPermaLink="true">https://nokaris.ca/en/insights/nestjs-angular-for-business-apps</guid>
      <description>For an e-commerce storefront I build headless. For the business application an SMB will keep for ten years, I choose NestJS and Angular — not out of fashion, but because the rigour they impose pays back in velocity over time. Conditions, trade-offs, and when I refuse this pairing.</description>
      <pubDate>Thu, 02 Jul 2026 00:00:00 GMT</pubDate>
      <category>devops</category>
    </item>
    <item>
      <title>Three mid-market AI projects that pay back within 90 days</title>
      <link>https://nokaris.ca/en/insights/three-ai-projects-that-pay-back</link>
      <guid isPermaLink="true">https://nokaris.ca/en/insights/three-ai-projects-that-pay-back</guid>
      <description>Counterpoint to the essay on cosmetic AI: three AI projects whose return is measurable within 90 days. Semantic search on a mature catalogue, support operations automation, copywriting under hard constraints. Conditions, metrics, method.</description>
      <pubDate>Thu, 04 Jun 2026 00:00:00 GMT</pubDate>
      <category>ai-applied</category>
    </item>
  </channel>
</rss>
