<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
<title>Elroot Labs Engineering Journal</title>
<link>https://elrootlabs.com/blog</link>
<description>Field notes on AI systems, local-first products and institutional software from the team building them at Elroot Labs.</description>
<language>en</language>
<atom:link href="https://elrootlabs.com/feed.xml" rel="self" type="application/rss+xml"/>
<lastBuildDate>Wed, 18 Nov 2026 00:00:00 GMT</lastBuildDate>
<item><title>Agentic AI, AI agents and workflows: when you need which</title><link>https://elrootlabs.com/blog/agentic-ai-vs-ai-agents</link><guid isPermaLink="true">https://elrootlabs.com/blog/agentic-ai-vs-ai-agents</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Agentic AI vs AI agents vs workflows: the sources disagree on the words. Ask who decides the next step, code or the model. Our own pipeline as the example.</description></item>
<item><title>AI agent observability: what to log on every tool call</title><link>https://elrootlabs.com/blog/ai-agent-observability</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-agent-observability</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We read the log of a six-hour session with a coding agent: 886 tool calls. What it recorded, what we had to rebuild, and the record we now want for every call.</description></item>
<item><title>AI agent permissions: a least-privilege checklist for tools and APIs</title><link>https://elrootlabs.com/blog/ai-agent-permissions</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-agent-permissions</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We mapped every permission our own writing agent holds. One limit was enforced in code. Most of the rest, including the key one, were lines in its prompt.</description></item>
<item><title>Is it safe to let an AI agent change production infrastructure? How we let ours do it</title><link>https://elrootlabs.com/blog/ai-agent-production-changes</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-agent-production-changes</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>When an AI agent can safely change live infrastructure, shown through a real change an agent made to our hosting, and the four it chose not to make.</description></item>
<item><title>AI agent tool failures: the five kinds we hit in one week, and how to recover from each</title><link>https://elrootlabs.com/blog/ai-agent-tool-failures</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-agent-tool-failures</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Seven tool failures from our own AI agents in one week, sorted into five kinds, with the recovery that fits each and the signal that tells them apart.</description></item>
<item><title>AI automation for small businesses: where to start without creating new risks</title><link>https://elrootlabs.com/blog/ai-automation-small-business</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-automation-small-business</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>AI automation for small business, one workflow at a time: how we automate our own small company with plain rules first and AI in one place, behind an approval.</description></item>
<item><title>AI trends for 2027, read from a year of search data</title><link>https://elrootlabs.com/blog/ai-trends-2027</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-trends-2027</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We put 98 technology searches on one scale for a year. Agents, coding and automation rose; LLM, RAG and MCP fell. What that suggests about AI trends in 2027.</description></item>
<item><title>Context engineering: what we put around the prompt, measured in words</title><link>https://elrootlabs.com/blog/context-engineering</link><guid isPermaLink="true">https://elrootlabs.com/blog/context-engineering</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Our daily writing agent could read about 107,000 words of files. Its playbook has it load about 20,000. What goes into its context, what stays out, and why.</description></item>
<item><title>Edge AI vs cloud AI: a cost calculation for 1,000 devices</title><link>https://elrootlabs.com/blog/edge-ai-vs-cloud-ai-cost</link><guid isPermaLink="true">https://elrootlabs.com/blog/edge-ai-vs-cloud-ai-cost</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>iot</category><description>Edge AI vs cloud AI, worked out with November 2026 prices: cloud wins for occasional text requests, edge wins by far for a camera checking every few seconds.</description></item>
<item><title>Google AI Mode and your website: what to check, from Google&apos;s own rules</title><link>https://elrootlabs.com/blog/google-ai-mode-website</link><guid isPermaLink="true">https://elrootlabs.com/blog/google-ai-mode-website</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>Google says AI Mode needs no special markup or files. We checked our own site against the rules it does set: four settings that can still keep you out.</description></item>
<item><title>Google Trends for keyword research: five traps we hit in our own data</title><link>https://elrootlabs.com/blog/google-trends-keyword-research</link><guid isPermaLink="true">https://elrootlabs.com/blog/google-trends-keyword-research</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>We used Google Trends twice in one week to pick blog topics. Five traps showed up in our data, from a sinking anchor term to queries that mean something else.</description></item>
<item><title>Hire AI developers or an AI development company? How to decide, and how to test them</title><link>https://elrootlabs.com/blog/hire-ai-developers</link><guid isPermaLink="true">https://elrootlabs.com/blog/hire-ai-developers</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>Should you hire AI developers or work with an AI development company? Who carries which risk, the four ways to engage a firm, and four tests for any AI hire.</description></item>
<item><title>Humanoid robot vs cobot: what each costs per hour of work, and the payback</title><link>https://elrootlabs.com/blog/humanoid-robot-vs-cobot</link><guid isPermaLink="true">https://elrootlabs.com/blog/humanoid-robot-vs-cobot</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>robotics</category><description>A humanoid robot from $13.5K or a cobot arm at $9,500? We worked out cost per working hour from the makers&apos; spec sheets, and the payback at Indian labour costs.</description></item>
<item><title>Setting up IndexNow on a Next.js site, including scheduled posts</title><link>https://elrootlabs.com/blog/indexnow-nextjs</link><guid isPermaLink="true">https://elrootlabs.com/blog/indexnow-nextjs</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>How we wired IndexNow into a prebuilt Next.js site with scheduled posts: one key file, one workflow that waits for the rebuild, and two mistakes we fixed.</description></item>
<item><title>Measuring LLM hallucination rate on your own documents</title><link>https://elrootlabs.com/blog/llm-hallucination-rate</link><guid isPermaLink="true">https://elrootlabs.com/blog/llm-hallucination-rate</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We asked two models questions our own passages could not answer. With a plain prompt, six of nine answers came from outside the passage. One line fixed it.</description></item>
<item><title>LLM structured output: no broken JSON in 80 calls, eight wrong answers</title><link>https://elrootlabs.com/blog/llm-structured-output</link><guid isPermaLink="true">https://elrootlabs.com/blog/llm-structured-output</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We ran 80 extraction calls with and without JSON schema enforcement. Every reply was valid JSON. All eight mistakes were in following the rules.</description></item>
<item><title>Post-quantum cryptography: what ML-KEM and ML-DSA change for an app</title><link>https://elrootlabs.com/blog/post-quantum-cryptography-ml-kem</link><guid isPermaLink="true">https://elrootlabs.com/blog/post-quantum-cryptography-ml-kem</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>security</category><description>We timed ML-KEM and ML-DSA against X25519, Ed25519 and RSA, and checked which sites already use post-quantum TLS. Speed was fine. Size is the problem.</description></item>
<item><title>Prompt versioning: changing a prompt without breaking production</title><link>https://elrootlabs.com/blog/prompt-versioning</link><guid isPermaLink="true">https://elrootlabs.com/blog/prompt-versioning</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Our writing agent&apos;s instructions changed five times in its first week. Each change shipped with the check that enforces it. What we version, and the gap.</description></item>
<item><title>A schema markup audit of our own site, before and after a rebuild</title><link>https://elrootlabs.com/blog/schema-markup-audit</link><guid isPermaLink="true">https://elrootlabs.com/blog/schema-markup-audit</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>We audited our own site&apos;s schema markup before and after a rebuild. Everything parsed both times; the move fixed one gap, kept one and added a new one.</description></item>
<item><title>Website security checklist for small businesses: what to fix first</title><link>https://elrootlabs.com/blog/small-business-website-security-checklist</link><guid isPermaLink="true">https://elrootlabs.com/blog/small-business-website-security-checklist</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>security</category><description>A website security checklist for small businesses in the order to fix it, with the one-line test for each item and the results when we ran it on our own site.</description></item>
<item><title>Vibe coding to production: what our AI-built website needed before it held up</title><link>https://elrootlabs.com/blog/vibe-coding-to-production</link><guid isPermaLink="true">https://elrootlabs.com/blog/vibe-coding-to-production</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>Most commits behind our website were co-written by an AI coding agent. These are the fixes it needed after the code worked, and the checks we run now.</description></item>
<item><title>Web5 vs Web3: what Web5 is, and why almost nobody searches for it</title><link>https://elrootlabs.com/blog/web5-vs-web3</link><guid isPermaLink="true">https://elrootlabs.com/blog/web5-vs-web3</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>Web5 vs Web3, with our own search data: Web5 scored zero on Google Trends. Its ideas live on as identity standards, and those are the part worth building on.</description></item>
<item><title>How we took our homepage from 51 to 88 on Google&apos;s mobile speed test</title><link>https://elrootlabs.com/blog/website-speed-core-web-vitals</link><guid isPermaLink="true">https://elrootlabs.com/blog/website-speed-core-web-vitals</guid><pubDate>Wed, 18 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>Five fixes that improved our Core Web Vitals in one day, each with its before and after number, in the order to check them on your own slow website.</description></item>
<item><title>How we built an AI blog pipeline, and what it checks before a post ships</title><link>https://elrootlabs.com/blog/ai-blog-pipeline</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-blog-pipeline</guid><pubDate>Tue, 17 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Our AI content pipeline writes only topics a person approved, then a 330-line script with no model in it rejects tells, unsourced numbers and unknown sources.</description></item>
<item><title>RAG chunking strategies, tested on the same document set</title><link>https://elrootlabs.com/blog/rag-chunking-strategies</link><guid isPermaLink="true">https://elrootlabs.com/blog/rag-chunking-strategies</guid><pubDate>Tue, 17 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>We ran eight RAG chunking strategies over our own blog with the same 24 questions. Labelling each chunk with its title and heading beat making chunks bigger.</description></item>
<item><title>Designing the approval step for AI agents: what the reviewer needs to see</title><link>https://elrootlabs.com/blog/agent-approval-step-design</link><guid isPermaLink="true">https://elrootlabs.com/blog/agent-approval-step-design</guid><pubDate>Sat, 14 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>What a human approval step for an AI agent has to show the reviewer, and the four ways the step fails quietly while still looking like oversight.</description></item>
<item><title>Denial of wallet: AI agent cost control before the bill arrives</title><link>https://elrootlabs.com/blog/ai-agent-cost-control</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-agent-cost-control</guid><pubDate>Sat, 14 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>What makes an AI agent run up a bill, from runaway loops to denial of wallet attacks, and the limits that stop it before the invoice does.</description></item>
<item><title>How to build an evaluation set for an AI agent from your own failure cases</title><link>https://elrootlabs.com/blog/agent-evaluation-test-set</link><guid isPermaLink="true">https://elrootlabs.com/blog/agent-evaluation-test-set</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Build an AI agent evaluation set from the failures that cost you trust, time or money. A worked example turns one incident into five cases and a release rule.</description></item>
<item><title>AI context platforms and the engineering problem underneath them</title><link>https://elrootlabs.com/blog/ai-context-platform</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-context-platform</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Context is a finite resource with diminishing returns. What an AI context platform has to solve, and the four techniques doing that work today.</description></item>
<item><title>ChatGPT search optimization: how to get your business cited in ChatGPT and Google AI answers</title><link>https://elrootlabs.com/blog/ai-search-optimization-chatgpt</link><guid isPermaLink="true">https://elrootlabs.com/blog/ai-search-optimization-chatgpt</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>ChatGPT search optimization comes down to the crawler you allow, snippet eligibility and honest markup. With a week of our own crawler data.</description></item>
<item><title>Cloud vs on-premise for small businesses in 2027</title><link>https://elrootlabs.com/blog/cloud-vs-on-premise-2027</link><guid isPermaLink="true">https://elrootlabs.com/blog/cloud-vs-on-premise-2027</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>Decide per system, not per company. What must keep working when the line drops, what the provider will never do for you, and what moving back actually costs.</description></item>
<item><title>Computer vision quality inspection: where it pays in manufacturing</title><link>https://elrootlabs.com/blog/computer-vision-quality-inspection</link><guid isPermaLink="true">https://elrootlabs.com/blog/computer-vision-quality-inspection</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Computer vision quality inspection pays when defects are visible and costly. How to choose a method when you have few or no photos of real defects.</description></item>
<item><title>Custom software vs off-the-shelf: what building our own website taught us</title><link>https://elrootlabs.com/blog/custom-software-vs-off-the-shelf</link><guid isPermaLink="true">https://elrootlabs.com/blog/custom-software-vs-off-the-shelf</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>Custom software vs off-the-shelf, from a company that built its own website and CMS: what it cost, what it bought, and a test for your own decision.</description></item>
<item><title>Drone development in India: the rules to design for from day one</title><link>https://elrootlabs.com/blog/drone-development-india</link><guid isPermaLink="true">https://elrootlabs.com/blog/drone-development-india</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>robotics</category><description>Drone development in India starts with the Drone Rules: weight class, type certificate, Digital Sky registration, airspace zones and radio approval.</description></item>
<item><title>Fine-tuning vs RAG: how to choose, and when a better prompt wins</title><link>https://elrootlabs.com/blog/fine-tuning-vs-rag</link><guid isPermaLink="true">https://elrootlabs.com/blog/fine-tuning-vs-rag</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>RAG fixes missing facts, fine-tuning fixes behaviour. We tested retrieval on our own blog and found where it really fails: the right page, the wrong paragraph.</description></item>
<item><title>Fixed price vs time and materials: choosing a software contract</title><link>https://elrootlabs.com/blog/fixed-price-vs-time-and-materials</link><guid isPermaLink="true">https://elrootlabs.com/blog/fixed-price-vs-time-and-materials</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>Fixed price suits a scope you can write down today. Time and materials suits work you will discover as you build. How to choose, and what moves the cost.</description></item>
<item><title>Guardian agents: what runtime oversight of AI agents actually requires</title><link>https://elrootlabs.com/blog/guardian-agents-runtime-oversight</link><guid isPermaLink="true">https://elrootlabs.com/blog/guardian-agents-runtime-oversight</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>What guardian agents are, why pre-deployment checks miss the failures that matter, and what the research says a supervisory layer has to watch at runtime.</description></item>
<item><title>IoT prototype development in India: from dev kit to an approved device</title><link>https://elrootlabs.com/blog/iot-prototype-development-india</link><guid isPermaLink="true">https://elrootlabs.com/blog/iot-prototype-development-india</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>iot</category><description>IoT prototype development in India in seven stages, with the WPC and BIS approvals built in from the first board so your design does not block the launch.</description></item>
<item><title>MCP server security: what the protocol does not check for you</title><link>https://elrootlabs.com/blog/mcp-server-security</link><guid isPermaLink="true">https://elrootlabs.com/blog/mcp-server-security</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>MCP server security starts from an uncomfortable fact. The protocol does not verify which agent is calling, so identity and scope are yours to add.</description></item>
<item><title>Why multi-agent systems fail: 14 failure modes and what they share</title><link>https://elrootlabs.com/blog/multi-agent-system-failures</link><guid isPermaLink="true">https://elrootlabs.com/blog/multi-agent-system-failures</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>Multi-agent system failures cluster into three categories and fourteen modes. What the research found in 1,600 traces, and why more agents rarely help.</description></item>
<item><title>MVP development: from idea to first users, with our own pipeline as the example</title><link>https://elrootlabs.com/blog/mvp-development</link><guid isPermaLink="true">https://elrootlabs.com/blog/mvp-development</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>MVP development in four phases, shown on a real example: the blog pipeline we built, what version one left out, and what real use forced us to change.</description></item>
<item><title>PCB layout mistakes to catch before you order the first board</title><link>https://elrootlabs.com/blog/pcb-layout-mistakes</link><guid isPermaLink="true">https://elrootlabs.com/blog/pcb-layout-mistakes</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>iot</category><description>Eight PCB layout mistakes that sink a first IoT board, each paired with the chip maker&apos;s own rule, so you can check the design before you pay for a run.</description></item>
<item><title>Running an LLM offline: what it takes and when it is worth it</title><link>https://elrootlabs.com/blog/run-llm-offline</link><guid isPermaLink="true">https://elrootlabs.com/blog/run-llm-offline</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>An offline LLM keeps data on the device and works without a network, at the cost of a smaller model. Five decisions settle whether it fits your app.</description></item>
<item><title>Sovereign AI: where your model runs and who can reach it</title><link>https://elrootlabs.com/blog/sovereign-ai</link><guid isPermaLink="true">https://elrootlabs.com/blog/sovereign-ai</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>Sovereign AI covers where a model runs, who operates the machines, and who can be compelled to grant access. What the deployment models actually differ on.</description></item>
<item><title>Who owns the code when you hire a software company in India?</title><link>https://elrootlabs.com/blog/who-owns-the-code-india</link><guid isPermaLink="true">https://elrootlabs.com/blog/who-owns-the-code-india</guid><pubDate>Fri, 13 Nov 2026 00:00:00 GMT</pubDate><category>buying-software</category><description>Under India&apos;s Copyright Act, a software company owns the code it writes for you unless your software development agreement assigns it in writing.</description></item>
<item><title>20 prompt-injection test cases to run against your AI agent before launch</title><link>https://elrootlabs.com/blog/prompt-injection-test-cases</link><guid isPermaLink="true">https://elrootlabs.com/blog/prompt-injection-test-cases</guid><pubDate>Sun, 01 Nov 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>A copyable prompt injection testing set for agents that read documents, email and web pages, grouped by where hostile text hides, each with its safe behaviour.</description></item>
<item><title>What safeguards should control an AI agent in production? The eight we run on our own</title><link>https://elrootlabs.com/blog/production-ai-agent-safety</link><guid isPermaLink="true">https://elrootlabs.com/blog/production-ai-agent-safety</guid><pubDate>Sat, 31 Oct 2026 00:00:00 GMT</pubDate><category>artificial-intelligence</category><description>The eight safeguards an AI agent needs in production, the order to add them, who signs off, and how each one works on the agent that runs our own publishing.</description></item>
<item><title>Local-first software is an operational decision</title><link>https://elrootlabs.com/blog/local-first-field-software</link><guid isPermaLink="true">https://elrootlabs.com/blog/local-first-field-software</guid><pubDate>Sun, 06 Sep 2026 00:00:00 GMT</pubDate><category>digital-infrastructure</category><description>When connectivity is uncertain, offline behaviour is not a fallback screen. It becomes part of the product’s core data model.</description></item>
<item><title>Build school software around the institution, not the dashboard</title><link>https://elrootlabs.com/blog/school-software-institution-first</link><guid isPermaLink="true">https://elrootlabs.com/blog/school-software-institution-first</guid><pubDate>Sun, 06 Sep 2026 00:00:00 GMT</pubDate><category>education-technology</category><description>Education systems succeed when they respect the daily work of teachers, administrators, students and families—not when they maximise feature count.</description></item>
</channel>
</rss>
