AI strategy & adoption

YouTube

How to choose the right home for your AI workflow

Dylan Davis May 13, 2026

---

AI strategy & adoption 9

Automation & tools 6

• AI defaults to building apps — that's usually the wrong choice
• 85–90% of workflows belong inside a project or skill, not deployed code
• Deploying an app triggers per-token API costs that subscriptions don't cover

Read more →

Automation & tools

YouTube

How to schedule AI tasks in Claude Cowork and Codex

Dylan Davis May 11, 2026

---

Automation & tools 9

Processes & SOPs 6

• AI output left inside the tool is output nobody reads.
• Automations with more than four steps fail far more often.
• A heartbeat and watchdog tell you when everything else breaks.

Read more →

AI strategy & adoption

YouTube

A four-step AI audit process for high-stakes outputs

Dylan Davis May 9, 2026

---

AI strategy & adoption 9

Processes & SOPs 5

• Almost-right AI answers are harder to catch than wrong ones
• Each audit step runs in a fresh AI conversation to prevent bias
• Four claim labels tell you exactly what to fix, remove, or review

Read more →

AI tools & automation

YouTube

How to use OpenAI Codex as a non-technical business user

Dylan Davis May 6, 2026

---

AI tools & automation 9

Automation & tools 6

• Codex brings the AI to your data — not the other way around.
• Every ChatGPT feature exists in Codex, but 2–3x more powerful.
• Five use cases that can save hours: files, browser, email, dashboards.

Read more →

AI tools & automation

YouTube

Seven signs it's time to move from browser AI to a desktop agent

Dylan Davis May 4, 2026

---

AI tools & automation 9

Automation & tools 6

• Long browser conversations degrade AI quality — fresh context fixes it
• Desktop agents can write their own rules, turning tools into compounding assets
• Scheduled autonomous tasks are only fully available on the desktop

Read more →

Prompt engineering

YouTube

How to update prompts for GPT-5.5 and modern AI models

Dylan Davis May 2, 2026

---

Prompt engineering 9

AI strategy & adoption 6

• Step-by-step mega-prompts actively limit what modern AI can do
• Binary success criteria help AI self-check before returning output
• Confident hallucinations require proof requirements, not just 'don't lie'

Read more →

AI tools & automation

YouTube

How to run autonomous Gmail drafting and weekly briefings with Claude

Dylan Davis April 29, 2026

---

AI tools & automation 9

Automation & tools 7

• Claude now loads only the tools it needs, not all at once.
• Train Claude on 300 past emails to draft replies in your voice.
• Schedule hourly inbox triage that runs while you sleep.

Read more →

AI tools & automation

YouTube

When to use Claude projects vs skills, and how to avoid overloading the AI

Dylan Davis April 25, 2026

---

AI tools & automation 9

Automation & tools 6

• Each active skill loads into every conversation, silently eating context.
• One project per activity, not per client — focus beats breadth.
• Capture skills from working conversations, not by writing them from scratch.

Read more →

AI tools & automation

YouTube

Breaking any role into AI-ready tasks: a five-step method

Dylan Davis April 22, 2026

---

AI tools & automation 9

Automation & tools 6

• Best AI users decompose tasks, not just write better prompts
• Use AI to interview you and extract your own process steps
• One activity per folder keeps AI focused and outputs higher quality

Read more →

Prompt engineering

YouTube

How to audit your Claude prompts after a model upgrade

Dylan Davis April 18, 2026

---

Prompt engineering 9

Processes & SOPs 6

• New models silently break setups that worked perfectly before
• Four checks catch tone, length, vagueness, and tool-skipping regressions
• Smarter models need fewer instructions — but far more precise ones

Read more →

AI tools & automation

YouTube

ChatGPT vs Claude vs Gemini: the hidden wrapper layer explained

Dylan Davis April 15, 2026

---

AI tools & automation 8

• The wrapper — not the model — determines what AI can actually do
• Poor memory management causes AI intelligence to degrade mid-task
• Desktop agents unlock file editing, cross-session memory, and system writes

Read more →

AI tools & automation

YouTube

Automating client data extraction with an AI funnel

Dylan Davis April 11, 2026

---

AI tools & automation 9

Prompt engineering 8

Automation & tools 6

• Wrong answers are 3x worse than blank — change AI's incentives.
• Force the AI to quote its source for every extracted value.
• Browser for testing; desktop agents for scale and self-improvement.

Read more →

AI tools & automation

YouTube

How to use sub-agents in Claude for parallel, independent tasks

Dylan Davis April 9, 2026

---

AI tools & automation 9

• Sub-agents run in parallel, keeping each worker's context fresh
• Only split tasks whose steps share no dependencies
• 3–4 sub-agents is the sweet spot; more burns usage fast

Read more →

Prompt engineering

YouTube

Trim your AI instructions to unlock better model performance

Dylan Davis April 7, 2026

---

Prompt engineering 9

• Bloated prompts actively restrain smarter models from performing well.
• 30–50% of rules in most prompts can be deleted without quality loss.
• A five-step quarterly detox removes stale, contradictory, and redundant rules.

Read more →

Prompt engineering

YouTube

How dictation closes the gap between your thinking and AI

Dylan Davis April 4, 2026

---

Prompt engineering 9

AI tools & automation 7

Productivity & habits 5

• Typing compresses your thoughts — dictation is 4x faster and richer
• The cringe factor fades in three days; most users never go back
• Three tactics: chunk, give AI a job, speak interview answers

Read more →