Phil Taul
Twenty years of networking, infrastructure, and business across a dozen industries. I build AI systems for real businesses, then I operate them. Everything published here runs in production first: the case studies below are live systems, not demos.
Build logs and playbooks from systems that actually run. No spam.
Three systems I built and operate today. Clients and identifying details are anonymized; the numbers are real.
A phone problem can live in any of six layers owned by five different vendors. This is how the ops stack from the earlier case study actually works: a diagnostic ladder that walks the whole call path from one conversation.
Read the case study
A watch-everything intelligence engine for one geographic market: public sources in, deduplicated and relevance-scored items out, a cited morning digest by 7 a.m. Adding a second market took a JSON file and zero code changes.
Read the case study
A one-person operation cannot afford a two-week rebuild after a dead machine. So both machines rebuild from one encrypted off-site copy, one script, one passphrase. Then the backups started failing on exactly one machine, and that was the point.
Read the case studyEvery project starts with a paid Workflow Roadmap Session, then takes one of three paths.
One 90-minute working session mapping your actual workflow, then a written roadmap: where automation pays off, what it takes to build, and what to skip.
I build the system and hand it over with documentation written so your team can run and maintain it without me.
I build it and keep operating it: monitoring, backups, and iteration as your workflow evolves. The same posture I use on my own systems.
A phone problem can live in any of six layers owned by five different vendors. This is how the ops stack from the earlier case study actually works: a diagnostic ladder that walks the whole call path from one conversation.
A watch-everything intelligence engine for one geographic market: public sources in, deduplicated and relevance-scored items out, a cited morning digest by 7 a.m. Adding a second market took a JSON file and zero code changes.
A one-person operation cannot afford a two-week rebuild after a dead machine. So both machines rebuild from one encrypted off-site copy, one script, one passphrase. Then the backups started failing on exactly one machine, and that was the point.
Every consultant hates timesheets, and every consultant leaks billable hours to bad memory. This system reconstructs the workday from evidence the tools already emit, and a human approves every entry before it becomes money.
A hyperlocal news platform: one weekly livestream becomes a week of short videos and articles, and a daily community brief drafts itself by 7:20 every morning. Humans still make every editorial call.
Per-minute AI clip tools make sense until you publish weekly. Then the math flips, and owning the pipeline wins on cost, quality, and control. Here's the math from my own show.
How AI tooling lets a single architect operate a multi-tenant cloud voice fleet, dozens of client networks, and a full security posture that used to need a team.
Phone systems taught me a standard: when it breaks, someone can't call 911. AI automation deserves the same discipline, and almost nobody builds it that way.
An auto dealer group was burning 20 to 30 staff hours a week transcribing vehicle titles onto state forms. Now an AI pipeline does it overnight, and a human only touches the titles that deserve a second look.
Every automation I run can draft an email, a reply, a report. None of them can hit send. That one design rule is why I can run AI against real business email without losing sleep.
Build logs, playbooks, and case studies from AI systems that actually run in production. Written by the one person who builds and operates all of them.