How to use AI beyond the basics and make it part of your everyday life.
The tools, your stack, and where Copilot fits
Get real work done faster
The main event — AI that takes action
Tools to explore and your action plan
AI is pattern recognition that got really, really good. You've been using it for years.
Knows what you want to watch next
Finishes your sentences as you type
Builds playlists around your taste
Suggests replies to your emails
Who uses Copilot every day? Who's tried ChatGPT, Claude, or Gemini?
Microsoft Copilot is a great start. It lives right inside the apps you already work in.
Drafts and summarizes email without leaving your inbox
Writes formulas and surfaces trends in your spreadsheets
Turns a document into a first-draft deck in seconds
Widest range of built-in tools. Image generation, voice mode, data analysis, browsing. Largest community.
Best at long documents, reasoning, nuanced writing, and coding. Follows instructions precisely. Least hallucinations.
Wired into Gmail, Docs, Drive, YouTube. Best AI video generation. If you live in Google's world, this is yours.
Most people use whatever loads by default. That's like driving a sports car in first gear.
| Platform | Quick Tasks | Hard Problems |
|---|---|---|
| ChatGPT | Default fast model | GPT-5.5 (paid) |
| Claude | Sonnet | Opus |
| Gemini | Flash | Pro |
The quality of what you get out depends entirely on what you put in. A few small changes make a huge difference.
"Write a professional 3-paragraph email to a client explaining a 2-day delay on their deliverable and proposing a new date" beats "write an email about a delay"
"You are a financial analyst. Review this budget and flag the three biggest risks..." gets sharper, expert-level answers than asking cold
"Here's an email I wrote that I like the tone of. Write the next one in the same style." It adapts to your voice.
Don't accept the first answer. Say "make it shorter," "more casual," "add a specific example." Treat it like a conversation, not a search.
Who has tried AI and felt like the results were just... meh? What did you ask it?
Most people use 10% of what these tools can do. Here's where they save you real time at work.
Drop in a 30-page report and get a summary with the key decisions and action items
Paste data and ask it to find trends or anomalies, or write the Excel formula you need
Draft replies in your voice; turn messy meeting notes into a clean follow-up plan
Put two contracts or vendor quotes side by side and ask it to compare and recommend
After a long planning meeting I pasted a wall of rough notes into Claude and asked for a structured summary: decisions made, owners, and next steps with dates. What used to take half an hour of cleanup came back in under a minute, ready to send to the team.
You can opt out of training, and business and enterprise tiers keep your data private and out of any model. Still, treat it like a smart coworker: share context, not passwords or credentials.
Free tiers are surprisingly good in 2026. You can get real work done without spending anything. Upgrade when you hit limits that actually bother you, not before.
AI can be confidently wrong. Always verify important facts. Think of it as a brilliant first draft, not a final answer. Claude tends to be most careful about accuracy.
AI replaces tasks, not people. The people who learn to use it will outperform those who don't. That's why you're here tonight.
An agent doesn't just answer questions. It takes action.
One exchange at a time. No memory. No follow-through.
Multiple steps. Uses tools. Makes decisions. Completes the task.
What's a repetitive task you wish someone else would just handle for you?
I had personal projects I'd wanted to build for years but kept hitting walls. I used AI to think through the parts I didn't understand, identify problems I hadn't anticipated, uncover hidden costs, and find workarounds. It became a thinking partner, not just a tool.
A chatbot waits for your next message. An agent takes a goal and runs with it.
You stay in the driver's seat, one prompt and one app at a time
Multiple steps, multiple apps, takes action, and hands back a finished result
A free, visual platform for building AI agents. No coding required.
Generous free plan to get started and build real automations
Drag-and-drop interface. See your entire workflow visually
Describe what you want in plain English. It builds the automation for you
It's the most approachable on-ramp to agents — describe the workflow in plain English with Gummie, and connect your tools with built-in MCP nodes.
Here's a receipt & invoice tracker built in Gumloop — four steps, no code.
Watches your inbox
Finds emails with receipts
Extracts vendor, amount, date
Logs them to a spreadsheet
A few that pay for themselves the first week. Each one is a workflow you'd describe in plain English.
Your calendar, top news, and priorities in one email each morning
Paste a link, get a post, an email draft, and key takeaways
New form submissions summarized and logged straight to your CRM
Track a topic and get a weekly digest of what's new
You describe what you want in plain English, and the AI builds it. No code, no technical background.
Same idea as Gumloop, more control. Open-source and self-hostable — worth knowing if you have IT or developer resources.
Run it on your own servers so sensitive data never leaves your walls
Hundreds of integrations and custom code when a workflow needs it
The choice when automations get complex or business-critical
A teammate on your own computer. Give it a goal and it works across your files, apps, and browser — then hands back a finished deliverable.
"Pull these five reports into one summary deck" — in plain English
Reads and edits files in folders you choose, fills forms, navigates the web
A real draft, spreadsheet, or document — not just instructions
Point it at a messy folder — it sorts, renames, and dedupes downloads and attachments
Hand it source files; get back a structured first-draft report or summary
Fills forms, pulls info from sites, handles repetitive web tasks for you
Reads and reconciles spreadsheets and CSVs, then writes up what it found
MCP is an open standard — think USB-C for AI. It plugs Claude into Google Drive, SharePoint, Slack, Notion, and thousands more, so Cowork works with the tools you already use. Plugins tailor it to a specific job, from finance to marketing.
Copilot is a great assistant. Cowork is closer to a coworker. Many teams use both — Copilot in the moment, Cowork for the heavy lifting.
| Microsoft Copilot | Claude Cowork | |
|---|---|---|
| Where it works | Inside one Office app | Across your files, apps & browser |
| How you use it | Prompt by prompt | Give a goal, it runs the steps |
| What you get | A suggestion to edit | A finished deliverable |
For advanced users — if you or your team write code, these are the most powerful agents out there. Totally fine if this one goes over your head.
A pair-programmer in your terminal. Understands a whole codebase and makes changes alongside you. Great for the important 20% — architecture and review.
A cloud agent inside ChatGPT. Give it a task, it works on its own and opens a pull request. Great for grinding through the routine 80%.
Choose a platform and go paid where it earns its keep
Rebuild one weekly chore as a Gumloop agent
Point it at one folder or report and see what it returns
Show one teammate what you built this week
The frontier: open-source agents you self-host — they live on your own machine, remember you, and reach you on the apps you already use. Two worth watching.
Self-hosted personal assistant reachable from WhatsApp, Telegram, Slack, and more. Huge community with 50+ integrations.
Open-source agent with a built-in learning loop — it builds skills from experience and gets sharper the longer it runs. Runs on a $5 server; your data stays yours.
Everything from tonight in one place. Model comparisons, starter prompts, tools, and links.
beyond-chatgpt-june.pages.dev
Thank you for spending your evening learning something new.
Rochester Brainery