<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>AI on Inquisitive M365</title><link>https://thomasdaly.net/category/ai/</link><description>Recent content in AI on Inquisitive M365</description><generator>Hugo -- gohugo.io</generator><language>en</language><copyright>© 2026 Tom Daly</copyright><lastBuildDate>Fri, 06 Nov 2026 03:00:00 -0500</lastBuildDate><atom:link href="https://thomasdaly.net/category/ai/index.xml" rel="self" type="application/rss+xml"/><item><title>OpenClaw, Hermes, Grok Bot: My Search for a Personal AI Assistant</title><link>https://thomasdaly.net/2026/11/06/agents-i-experimented-with/</link><pubDate>Fri, 06 Nov 2026 03:00:00 -0500</pubDate><guid>https://thomasdaly.net/2026/11/06/agents-i-experimented-with/</guid><description>I wanted a personal assistant I could talk to, to keep up with my email, calendar and projects. It took three AI agents, two extra computers, and a lot of my own connectors to get there.</description></item><item><title>Finding Lease Data in Giant Scanned Packets</title><link>https://thomasdaly.net/2026/11/05/archive-lease-extraction-giant-packets/</link><pubDate>Thu, 05 Nov 2026 00:00:00 +0000</pubDate><guid>https://thomasdaly.net/2026/11/05/archive-lease-extraction-giant-packets/</guid><description>Giant scanned packets of leases, rental agreements, cancellations and letters. How we pulled the files from SharePoint, ran OCR on a local machine, read them with a cheap AI model in Microsoft Foundry, and only paid for a stronger model when a simple check failed.</description></item><item><title>Web Parts Are Coming</title><link>https://thomasdaly.net/2026/11/05/web-parts-are-coming/</link><pubDate>Thu, 05 Nov 2026 00:00:00 +0000</pubDate><guid>https://thomasdaly.net/2026/11/05/web-parts-are-coming/</guid><description>I&amp;rsquo;m pulling together the SharePoint web parts I&amp;rsquo;ve built over my career, plus the ones people keep asking for. Here&amp;rsquo;s what&amp;rsquo;s coming, and how AI lets me build them properly and still support them.</description></item><item><title>Automated Lease Routing with Microsoft Foundry, Azure Functions, and SharePoint</title><link>https://thomasdaly.net/2026/08/18/azure-ai-document-routing-sharepoint/</link><pubDate>Tue, 18 Aug 2026 00:00:00 +0000</pubDate><guid>https://thomasdaly.net/2026/08/18/azure-ai-document-routing-sharepoint/</guid><description>How I used Azure AI to read lease documents and file them in the right SharePoint folder, and the surprises that didn&amp;rsquo;t show up until real documents started flowing through it.</description></item><item><title>Why We Rebuilt a Power App as a Web App (And How AI Helped Us Ship It Fast)</title><link>https://thomasdaly.net/2026/03/05/power-app-to-web-app-migration/</link><pubDate>Fri, 06 Mar 2026 00:00:00 +0000</pubDate><guid>https://thomasdaly.net/2026/03/05/power-app-to-web-app-migration/</guid><description>A boutique camera parts dealer outgrew their Power App. Here&amp;rsquo;s why we moved to a custom web app, the problems that forced the switch, and how AI-assisted development made it practical for a small team.</description></item><item><title>Building a Two-Stage AI Pipeline for Invoice Processing with AWS Textract and Amazon Bedrock</title><link>https://thomasdaly.net/2025/03/04/textract-plus-bedrock-two-stage-ai-pipeline/</link><pubDate>Wed, 05 Mar 2025 00:00:00 +0000</pubDate><guid>https://thomasdaly.net/2025/03/04/textract-plus-bedrock-two-stage-ai-pipeline/</guid><description>How I built a serverless pipeline using AWS Textract for OCR and Amazon Bedrock for AI normalization to extract structured data from invoices across hundreds of different vendor formats.</description></item></channel></rss>