Professor Glitch
@pro.glitch • 0 subscribers
Videos
How to price AI automation (the math, the packages, and what to say on the call) How to price an AI automation project for a small business, with the real math. 1. Price the problem, not your time. Example: a heating and air company's office loses 24 hours a week to follow-ups ($42,000 a year) and about one job a month to quotes nobody chased ($48,000). The problem costs about $90,000 a year. 2. The 5× rule: the client should get back at least five times what they pay in year one. Value sets the ceiling, your costs set the floor (double your gut hours). 3. Sell it in three parts: map it (paid, $2–3K for a small company), build it (fixed price, never hourly), run it (monthly, e.g. $800). 4. Send three options, not one: $6,500 / $12,000 / $22,000. Most clients pick the middle. 5. The client owns the accounts and pays AI usage directly. 6. On the call, ask instead of pitching, and let them say the value number. Say your price, then stop talking. No clients yet? Do the first one at your job: time it before and after. That's your case study. #aiautomation #aiemployee #aiagents #aiconsulting
Professor Glitch44,200 просмотров • 12 часов назад
Building an app with AI? Learn these 4 parts first Want to build an app with AI? Learn the four parts first: the front end, the back end, the database, and where AI fits in a modern app. When you ask ChatGPT or Claude to build you an app (vibe coding, no code), these four parts are what you're actually describing. Explained for complete beginners. We follow one customer message through an online coffee store app, start to finish: 1. Front end: everything the customer sees and taps. The screens, the buttons, the text box. It doesn't know prices or stock. It shows things and passes messages along. 2. Back end: runs on a server the customer never sees. This is where the business rules live: login, stock, prices, payments. The front end asks, the back end decides. 3. Database: where everything is saved, like a giant spreadsheet that never closes. Products, prices, customers, orders. Only the back end opens it, and it checks who's asking. 4. AI in the middle: the back end calls the AI model only for jobs that need understanding. Job one, turn a plain sentence ("a gift for my dad, he drinks it black, under $40") into a search. Job two, write the answer using only the real products the database sent back. The rule for building reliable AI apps: AI handles the words, the back end handles the facts and the money. Never let the AI make up prices or stock. Whole app in four lines: the front end shows, the back end decides, the database remembers, the AI understands and writes. If you want to go from understanding the four parts to building your own apps with AI, I teach it step by step, from complete beginner to real systems. Check my profile. #vibecoding#buildwithai#claude#claudecode#aiagents #
Professor Glitch61,200 просмотров • 1 день назад
How to build a complete app with AI (no code, the whole picture) How to build a complete app with AI when you don't know how to code: login, a database, photos, invoices, and a real link your customers open on their phone. We follow one client, Priya, through a client portal for a kitchen remodeler. She logs in and sees her project's stage, her photos and her invoice. Every step answers a real beginner question about how apps work. 1. Your workshop: VS Code with the Claude Code extension, an empty folder 2. Describe the app in plain English (plus three technical lines you can copy) 3. The front end, and why localhost means only you can see it 4. The database: the refresh test, and what tables are 5. Login: why your password is stored only as a hash, and what a session cookie does 6. The one rule every app needs: each customer sees only their own data, proven with a test 7. Move the database online: from a file on your laptop to a hosted Postgres database on Supabase, with secret keys in a private .env.local file 8. GitHub: put your code online in a private repository 9. Vercel: deploy, add environment variables, and get a live link (honest costs: Vercel Pro for a business, Supabase Pro) 10. Your own domain with DNS and HTTPS, then a security audit before real clients log in Works for any small business app: client portals, booking apps, customer dashboards, internal tools. AI app builder, vibe coding, no-code app, full stack app with Claude Code. I have an entire course on building apps with AI from zero in my community. If you're interested, the link is in my profile. #vibecoding#aiapp #claudecode#aiagent
Professor Glitch98,000 просмотров • 3 дней назад
How to build a second brain that updates itself (like a librarian) How to build a second brain that updates itself: an AI knowledge base for your business that knows your clients, your prices and how you work, and keeps itself up to date every time you give it something new. The setup, in plain words: 1. One folder on your computer with two shelves: an inbox where you dump everything (call notes, emails, PDFs, voice memos, ideas) and the brain, with clean pages, one per client, project, pricing, and one about you. 2. One rules file that turns Claude into your librarian: never change an original, every fact points back to its source, mark old facts outdated instead of erasing them, and link everything related. 3. File it: drop a note in, and the AI updates every page it affects (often five or six) and links them back to the note. 4. When something changes, the brain updates itself. A client cuts their budget from $8,000 to $6,000 on a call: the client page, the timeline and the revenue page all update, and the old number stays, marked outdated, with the date and the source. 5. Ask it anything and it answers from your own pages, with the source. 6. Once a week, it checks the shelves: pages that disagree, facts that went stale, pages nothing links to. It's all plain text (markdown) files on your computer, and you own them. Open the folder in Obsidian to watch the connections grow. The whole job in four moves: you drop, it files, you ask, it checks. If you want to build yours, I teach it file by file, plus full courses from complete beginner to building real AI systems. Check my profile. #secondbrain #claude #aiagents #claudecode #professorglitch
Professor Glitch10,200 просмотров • 1 день назад
You have exactly four months to change how you work forever with AI and Automation Four months. September, October, November, December. That's exactly enough time to go from wherever you are to running AI automations and your first AI employee. But the whole thing hinges on one week — this one. SEPTEMBER — foundations + your first automation. Learn the words fast, then build one small thing. By September 30, for the first time in your life, something you built runs while you sleep. OCTOBER — AI in the middle. Your automations start to think: the email comes in, it reads it, decides, drafts, hands you only the tricky ones. By Halloween, the busywork is off your plate — a machine handles it, in your voice, all day. NOVEMBER — your first real AI employee. Connect it to your tools, write its instructions, give it memory, set the brake so the risky stuff waits for your yes. By November 30, you have an employee working one real job, every single day, getting sharper every time you correct it. DECEMBER — multiply. The second takes a fraction of the effort. And then New Year's Eve arrives: everyone around you is writing resolutions… and you're already running. Here's the catch, and it's the only one: four months is exactly enough, with nothing to spare. So the version of you holding all of this at midnight doesn't start in October — they start this week. The four months are the easy part. This week is the decision. That's the whole difference between the two people you could be at midnight on the 31st: the one who started, and the one who meant to. Do it right — in order, from your first automation, with a group starting the exact same week. Get in this week and start September on foundations. It's all in my profile. #aiautomation #aiagent
Professor Glitch360,300 просмотров • 2 месяцев назад
How to build an AI receptionist that answers your business phone (books jobs, knows your customers) The whole build, on a real phone number. A Retell voice agent with your own Google Docs as its knowledge → books jobs in Cal.com → exact quotes from a small calculator (the AI never adds prices) → n8n looks up the caller in your Google Sheet before she says a word ("Hi Jen") and checks where the cleaner is mid-call → warm transfer to a person for complaints and damage → office leads into HubSpot → a summary of every call in Slack and your sheet. Tested on 21 hard simulated callers: 17 passed on the live version, and the misses go into the next one. About 13 cents a minute plus $2/month for the number. The full AI receptionist course, with every workflow, is coming to my community soon. #aiagents #aireceptionist #aiemployee
Professor Glitch31,000 просмотров • 7 дней назад
How to build an AI front desk for your info@ inbox in Microsoft 365 (no $30 Copilot seats) Built for real in a Microsoft 365 lab: shared mailbox → Power Automate flow with cheap checks first (own mail, out-of-office, duplicates, mail loops) → a Copilot Studio agent that answers only from your SharePoint files → a Teams approval card → the reply goes out from info@ in the same thread. Plus a chase if nothing's answered and a morning digest. Tested on 20 emails: 18/20 routed right on the first run, 8/8 after two instruction fixes, zero wrong replies reached a customer. The listening flow costs nothing; the AI costs a few cents per drafted email. The build, step by step: 1. Microsoft 365 admin center: create a shared mailbox called info@ (no licence needed) and give yourself access. 2. Make a Teams team for the front desk. In its SharePoint, add your own documents (prices, booking and cancellation policy, vendor rules, FAQ, directions) plus two lists: who handles what, and an inbox log. 3. Copilot Studio: create an agent on a fast, standard model (not a reasoning one), point its knowledge at those files, and tell it to answer only from the files, name the file it used, and hand dates, money, complaints, gate codes, invoices and anything it can't find to a person. 4. Power Automate: trigger on a new email in the shared mailbox, run the cheap checks before any AI, log the rest and send it to the agent. 5. Teams: the flow posts a card with the draft, the reason and the source file. Approve and the reply goes out from info@ in the same thread. 6. Nothing gets forgotten: a second flow chases anything unanswered after 24 hours, and a third posts a morning digest. Two setup details that stop most people: a company data policy can block the agent from reading your own files, and the agent needs generative orchestration turned on or it ignores its instructions. Works for venues, clinics, agencies, trades, offices, any small business with an info@ inbox and Microsoft 365. #copilotstudio #aiautomation #powerautomate #claude #microsoft365
Professor Glitch22,600 просмотров • 9 дней назад
How to build an AI invoice clerk inside Microsoft 365 (Copilot Studio + Teams) Every supplier invoice that lands in Outlook gets filed in SharePoint, read by AI, checked against your payment rules, and either approved, sent to a manager in Teams, or stopped. Built for real in Copilot Studio: 1. Outlook trigger, only emails with attachments 2. File every PDF in SharePoint 3. Extract reads the invoice (plus one field that looks for changed bank details) 4. Classify filters out junk attachments 5. An agent checks your payment rules and explains every decision 6. Big or risky invoices wait for one tap in Teams 7. Every decision lands in one log, with the reason The AI never pays anything big on its own and never touches a bank change. A person decides those. And this is one desk. Add a three-way match, supplier emails, a bank-change freeze, a Friday payment run and an agent your team can ask, and it's a whole finance department on AI. I'm adding an entire course on the Microsoft ecosystem to my community: complete AI departments, built for real, step by step. #microsoftcopilot #aiautomation #aiagents #microsoftteams
Professor Glitch20,800 просмотров • 15 дней назад
The first 5 AI employee to make #aiemployee #aiagent #aiautomation
Professor Glitch46,200 просмотров • 1 месяц назад
AI feels like too much? Here's what to learn first. #claude #aiagent #aiemployee #aiautomation
Professor Glitch10,500 просмотров • 11 дней назад
The Complete AI & Automation Roadmap (Beginner → Expert) #professorglitch #aiautomation #airoadmap #aiagent #rag #learnai
Professor Glitch170,600 просмотров • 11 месяцев назад
How to make money with AI by starting a real business — in a matter of weeks (the honest version) The exact path to making money with AI by starting a real business — no bank-account screenshots, no rich-by-Friday. Here's the whole thing: FIRST, THE CHOICE → there isn't one way, there are two — and most people who quit picked the wrong path, not the wrong dream. Path 1: the solutions business — find businesses with a boring painful problem (missed follow-ups, unanswered reviews, chased invoices), get on calls, sell the fix. Path 2: the product — build one simple thing that kills one pain, make videos about it, sell it to anyone. Scalable, no sales calls. BUILD THE DEMO → before you have a customer. Some people take a lesson, build it on their OWN use case, and in a day have a working demo. It stacks: walk into any call with a library — "oh, I've got something for that. Want to see it run?" SOLUTION FIRST, TOOLS SECOND → customers don't buy tools, they buy solutions. See the obstacle, design the system, then the technology. A tool without a solution behind it builds something that breaks. TWO ENGINES → outreach (not spraying — AI wires the tools that find the RIGHT people and personalize) or content (talk daily about the problem you solve; they reach out to you). CONSISTENCY, DONE RIGHT → two people at the same gym every day: one grinds for years eating trash — barely changes. One measures everything, adjusts, iterates — gets there every time. Same gym, same effort: one understands the system. Business is exactly this — the game pays out in month two, and most people quit in week three. MEASURE → a simple sheet: what you sent, posted, who replied. Dead numbers after a month or two = diagnose: the product, the message, or the people. Fix one, run again. #aiagents #makemoneywithai #aiautomation
Professor Glitch35,600 просмотров • 3 месяцев назад
How to Build second Brain My entire second brain is on a hard drive — not in ChatGPT, not in Claude, not in anybody's app. It's folders and files. I can unplug it, hand it to a completely different AI tomorrow, and it still knows everything about me and my business. The whole build: • raw/ — yours. Every note, meeting transcript, voice memo, screenshot. You dump, you don't organise. • wiki/ — the AI's. It reads raw, works out what matters, writes clean pages, and links them together. That's the part that matters — most "chat with your documents" tools just find chunks that look similar and hope. They have no idea two documents are related. Yours does, so it walks from one page to the ones it's tied to and picks up the context it was missing. • index.md — the map. Without it, the AI reads your entire brain every time you ask one question. • log.md — the memory. So a new document six weeks later updates the old page instead of duplicating it. • A hook or a schedule keeps it running without you. And here's why it's the foundation: when you build AI employees later — the one answering your email, the one going through your numbers — they don't start from zero. They feed from this. They already know how you talk, what you charge, who your customers are. Because the best model on earth still doesn't know your business. This is the thing that closes that gap. Full step-by-step build + templates this week. Check my profile. #aiagent #aiautomation #secondbrain
Professor Glitch20,200 просмотров • 3 месяцев назад
Больше нет контента для загрузки