Used in this story
Singapore
For a family move to Singapore, he asked Grok Bot to find furnished three-bedroom rentals for three months. It put the options in a Google Sheet. After his wife picked five, it used his calendar to arrange visits and helped with the lease talks. He told this in a conference talk, and he works on the product. The proof link is a timestamped summary of the talk.
Actual tools: Grok Bot, Google Sheets, Gmail, Calendar
Result as claimed: It compiled about 35 options in a Google Sheet and helped negotiate leases for two properties.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Philippines
Mikes Sta. Ana built this system for Coach John. It sends check-ins in his voice across a 30-day programme, scores leads, onboards new clients with a payment link and drafts social posts for him to approve. The source is the builder's own case study, and the coach is not quoted.
Actual tools: n8n, Telegram Bot API, Supabase, PayMongo, Claude, GPT-4o-mini
Result as claimed: Daily time on operations went from 45+ minutes to 3 minutes. Onboarding after a sale went from 1 to 3 days to under 60 seconds. Active client capacity went from 15 to 50+. AI API cost is under $1/month.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Indonesia
Ko Andi runs a kos-kosan (boarding house) on his own. His agent reads guest check-in reports from a Telegram group, enters them into Google Sheets, matches them against bank transfers and builds the report. The source is a blog post in Indonesian from KomuniTech, the community that sells the course and membership he joined.
Actual tools: Telegram, Google Sheets
Result as claimed: Time and effort for this workflow dropped 60-70%, a figure the source attributes to Ko Andi himself.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Spain
He runs a small car-wash equipment service company and is not a programmer. With Claude Code he built an ERP that replaced the commercial package they used for years, plus agents that read the mailbox, create service calls, draft replies and prepare supplier orders. Anything with consequences goes to his phone on Telegram for approval.
Actual tools: Claude Code, MCP server, Telegram, Local models
Result as claimed: "Our own ERP, in production, replacing the commercial package we used for years." He reviews "twenty decisions a day from the phone in a few minutes".
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
He could not work out the sell-and-buy cash flow in Excel, so he built a calculator with Claude Code, mostly from his phone. It shows BSD, CPF refund, SSD, mortgage stress tests and a month-by-month timeline on one screen. He published it as open source.
Actual tools: Claude Code
Result as claimed: "Built a Singapore property upgrade calculator, mostly on my phone, with Claude Code." He notes it gives estimates only.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
NOVA, his custom agent built with Notion AI's custom instructions, reads his Notion workspace for the week and writes a structured weekly reflection with outputs, outcomes, an activity map, lessons and an infographic. For long jobs he adds "Apply COMPASS", a quality gate that makes it check and correct each step before moving on.
Actual tools: Notion AI
Result as claimed: Weekly reflection: "For years it took half a day and still felt incomplete." Now: "The whole thing takes 15 to 20 minutes to generate." Long jobs: "My highest record was 50 tasks in one prompt."
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Green-Tech Egg Industries gets around 120 recurring customer orders a day as WhatsApp text and voice messages. Staff used to read or listen to each one and key it into AutoCount, the company's ERP, at about five minutes an order. MEVO Anna, an AI sales-admin agent, now reads each order, confirms the items and quantities with the customer, and enters only confirmed orders into AutoCount. The source is MEVO's own news page, so the figures are not independently checked.
Actual tools: MEVO Anna, WhatsApp, AutoCount
Result as claimed: Green-Tech has achieved approximately "80% customer adoption", "8 man-hours saved per day" and "80% reduction in delivery errors".
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Philippines
Julie's Franchise Corporation runs the franchise network for more than 600 Julie's bakeshops across the Philippines. Its partner Computrade Technology Philippines built a procurement agent on Claude Sonnet 5 and Amazon Bedrock AgentCore that reads each supplier bid, weighs pricing and terms, and drafts a side-by-side recommendation. Procurement officers review, revise and approve every decision. The source is Computrade's own case study.
Actual tools: Claude Sonnet 5, Amazon Bedrock AgentCore, AWS
Result as claimed: "There were less time to compare 3+ quotations. From 2 to 3 hours down to 5 to 10 mins using the new AI application," said Crisanta Tatad, National Procurement Head. The case study says supplier decision time "has fallen by 95% with no additional procurement headcount".
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United Kingdom
Tristan Watson runs Fieldtrip Supply Co., an engraved outdoor gear brand in Tynemouth, England, on his own. Claude Code is connected to fourteen systems including Shopify, Meta Ads, Google Ads, Xero and Gmail. Each morning it drafts personal thank-you emails for overnight orders, sorts the fulfilment queue, flags low stock and reviews ad spend against his margin targets. It also reconciles his bank account against Xero and triages his inboxes. He says Claude Code built these systems from his descriptions and that its ad creative still needs heavy human work.
Actual tools: Claude Code, Shopify, Meta Ads, Google Ads, Xero, Gmail, Google Search Console
Result as claimed: "The business has grown nearly 4x in the last twelve months, but it’s still just me." "One person, running a business that turned over £300,000 last year." His Monday sales and profit review "used to be at least an hour’s work every Monday morning" and is now a live dashboard he can query.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Erick Villalobos acquired Evermove Engineering, a 52-year-old Singapore industrial supplies company, with 12 years of transaction records, about 1,000 customer accounts and no data team. He described the problem to Manus, an AI agent, in plain English: which customers to call, when and in what order. Manus offered several options, including a custom CRM. After a week of daily use and fixes it was producing the sales team's weekly call schedule, and he keeps improving it each month. The account is written by NUS Business School, where he took an AI course.
Actual tools: Manus
Result as claimed: "Two hours later, a working version of the CRM was running on Evermove’s real company data." "Total budget: zero." A CRM at a previous company "cost S$25,000 and was abandoned on the first day".
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Malaysia
Eunice Yin Ern runs OTSATE, a small lemongrass sauce brand in Malaysia, and does the cooking and the numbers herself. Every Monday she tells Claude Code last week's sales, ingredient and packaging spend, cash in the bank and the decision she faces. It writes the row in a Google Sheet, works out revenue per day, food cost percentage and weeks of fixed costs covered, and a script it wrote sorts her bank statement PDFs into categories for her to confirm. When it costed a new product with supermarket prices, she added a standing rule that it must never estimate her costs.
Actual tools: Claude Code, Google Sheets
Result as claimed: "What I stopped doing is opening the spreadsheet to type. What I do with that time is read it. Ten minutes, every Monday, and I actually look." On the new product, the AI said contribution of "12 to 19 per cent", but "With my real prices the contribution was 61 per cent."
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Raine Goh founded Business Buddy, a Singapore digital solutions firm, in 2013. Her team uses AI for quotations, routine reports, basic troubleshooting, code and content drafts, and recently built an AI customer service agent. She retrained her writers as prompt writers and moved consultants toward client work.
Actual tools: ChatGPT
Result as claimed: "Take quotations, for instance," she says. "What used to take my consultants over a week now takes about an hour."
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Spain
VJ Roworth and Adrian Beadell run Leonard's Lookout, a four bedroom holiday rental in a cave house in Spain. Their ChatGPT agent, which they call Scout, replies to Airbnb bookings, cross checks calendars across booking platforms to avoid double bookings and keeps a plant care schedule. When they asked how to get more visitors, it suggested press, found a Forbes contributor, drafted the pitch and prepared talking points for the interview.
Actual tools: ChatGPT
Result as claimed: Roworth says the Forbes story would not have happened without Scout. "In terms of saving money, we can't afford a business or an admin assistant or a marketing guru, and Scout's been all of those things for us," she said.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Gavin Sim runs an AI education programme and helps run a coaching company in Singapore. Many of his short videos, including client brand campaigns, are made with an AI avatar of himself in Google Flow. He talks through an idea with an AI agent, which returns a storyboard, generates the clips and assembles them. A storyboard template fixed early problems like his shirt changing colour between clips.
Actual tools: Google Flow
Result as claimed: "A video that used to need a shoot day, a crew and an editor now comes together at my desk." He also writes that if he put in 40 hours a week last year, "those same 40 hours now produce what used to take 120 or more."
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Indonesia
Lucas Zhu, an ex-McKinsey founder with no engineering background, ran a group of AI agents like a small team to build a point of sale app for the Frollie cookie booth in Block M, South Jakarta. It covers staff sign in, stock, sales, QR and bank transfer payments, receipts and manager approvals on Telegram. He wrote his own review and testing workflows because, he says, the agents did not catch everything.
Actual tools: AI coding agents (not named in source)
Result as claimed: "Fifteen days ago this was a planning doc. Today it takes real money at our cookie booth in Jakarta, and I hired no engineers to get there." His deck lists idea to live in 12 days, 500+ automated tests and taking real money on day 9.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Philippines
Angeli Soriano-Gojo, a former banker, has run StonedMNL, a crystal and astrology business, for 10 years. She had Gemini help write prompts, saved them as a framework in a PDF, and runs it with her crystal directory to match each client's sun, moon and rising signs and concerns with crystals. She also uses Gemini to change photo backgrounds and lighting, draft captions and emails, organise consultation notes and keep her crystal directory consistent. She checks and edits the output herself.
Actual tools: Gemini
Result as claimed: "This process used to take 30 minutes and a lot of times, clients don't have that time. Now, it just takes three minutes to compute the client's sun sign, moon sign, and rising sign," she said. She could not give percentages or figures, but said business has improved.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Alan Tan is the third generation owner and CEO of HarriAnns Nonya Table, which supplies kueh to hotels and runs eight cafes. In March 2025 a hotel's 2,000 kueh order was missed because pages of its purchase order were not printed. Tan had Singapore consultancy Knovel Engineering build an AI platform that reads purchase orders from email, recognises the naming variations of over 120 products, tallies the next day's production and flags unusual ordering patterns. It went live in June.
Actual tools: Custom AI ordering platform (Knovel Engineering)
Result as claimed: "That reduced a whole day's work for my admin staff to just two hours. They just have to spend time checking the AI's output," said Tan. The system cost less than $50,000 to develop, and an Enterprise Development Grant covered half.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Lin Enqi took over her family's neighbourhood grocery shop at 18. When the shop moved to a new point of sale system, she used ChatGPT to categorise more than 8,000 products. She asked it for ways to lift sales without hurting margins and it suggested vouchers tied to a minimum spend. With Nanyang Polytechnic's Centre for Applied AI she built a custom GPT that suggests festive and seasonal promotions, and she uses ChatGPT to keep up with retail trends.
Actual tools: ChatGPT, Custom GPT
Result as claimed: CNA's Money Mind says she uses AI to save time, boost sales and keep the shop profitable while paying her own university fees. Reports citing CNA say the product categorisation took about two days with ChatGPT, compared with nearly a week by hand, and that the shop carried more than S$30,000 in debt when she took over. In her own LinkedIn post she says ChatGPT helped with "organising over 8,000 products" and repetitive tasks.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Dhawal Shah runs 2Stallions, a regional digital marketing agency. For most of 2026 he has run his own marketing through Claude Code, connected to his tools through their APIs. On his own site it handles the website, SEO, analytics, newsletter, social and video. Each Monday he asks one question about what changed across the site, search, ads and newsletter and gets an answer in about a minute. On the agency side it only runs ad reporting across five platforms, after a privacy and terms of service review. He signs off anything that spends money.
Actual tools: Claude Code
Result as claimed: "The recurring work of running every channel used to take me about 25 hours a week. It now takes about five." A weekly cross platform ad report went from four to six hours to about 30 minutes, and a LinkedIn content pack from three or four hours to about 30 minutes. The stack costs about US$250 a month on AI.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Thailand
Phinyaphat Junsoton, known as Khun Nui, is a Data Entry Officer and PR representative at Phichit Drinking Co., Ltd. She handles scheduling and employee data for more than 200 staff across two shifts and six production lines. Instead of searching for and testing Excel formulas by trial and error, she explains the rules of the monthly schedule to Microsoft Copilot and asks it to build the formulas she needs. She also uses Copilot for market research and planning toward a hydroponic farming side income.
Actual tools: Microsoft Copilot, Excel
Result as claimed: Microsoft reports that the monthly shift schedule took "at least three full days" before, and "could now be done in about 30 minutes."
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United States
Carrol Chang, CEO of Andela, is not an engineer. She asked AI to teach her step by step and built agents for her family. A sports agent maps the commute to each of her kids' games and puts travel time on the calendar, and it sent about 60 calendar invites for soccer games and the trips there and back. A school agent reads school emails, adds closures and relevant events to the calendar, ignores events for the other language track, and passes on emails from teachers. A meal agent suggests menus that meet the family's allergy needs and makes a shopping list.
Actual tools: AI agents (models not named in source)
Result as claimed: "It was about five to eight hours a week of mental labor... Now, I probably spend two hours a week doing this stuff and checking on the agents."
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Keith Teo wanted to check his own flat's value without handing his phone number to property portals. He used Claude Code to write a data pipeline and page templates that pull more than 227,000 HDB resale transactions from the government's open data portal and generate a page for every block. The site, CheckHowMuch.sg, is built with Astro and hosted on Cloudflare Pages. He then fixed 16 problems found after launch, from missing flat types to a valuation formula users said was too low.
Actual tools: Claude Code, Astro, Cloudflare Pages
Result as claimed: "Two days later, I had CheckHowMuch.sg." The site has 9,700+ pages covering every HDB block, with no login or phone number needed, and "costs $0 per month to run."
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United States
Matt Murphy sells trading cards on eBay. He built a pipeline with Codex that wakes up each night and works through a batch of 50 photographed cards: it identifies each card, researches what it sold for and drafts the listing. He reads and publishes every draft himself, and it stops and asks when it isn't sure.
Actual tools: Codex, OpenAI models
Result as claimed: By hand each card took him 5 to 10 minutes. The pipeline's median batch run of 50 cards cost $42.39 at API list prices, about $0.85 a card, plus about 30 seconds of his review. Valuing his time at $150 an hour, he puts the saving at $10.40 to $22.90 a card and $520 to $1,145 a batch, so the $1,200 build paid for itself by the third batch.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Kopi Near Me runs about 200 coffee and tea machines across Singapore. Its small team works with AI agents to quote clients, schedule technicians, plan replenishments and keep operations records. The scheduling work for the fleet is now done by the sales admin and ops manager with AI agents.
Actual tools: AI agents (not named in the post)
Result as claimed: "Scheduling for our fleet of 200 machines islandwide used to take 10 hours a week. Today, our sales admin and ops manager, working with AI agents, get it done in 2." Kopi Near Me also reports a 70% reduction in manual enquiry processing, 25% fewer unnecessary technician trips, 40% fewer emergency replenishments and an estimated 10 to 15% better sales conversion.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
David Bressler runs Formula Bot, an AI data analysis app, with Bubble for the front end and n8n for the backend. One orchestration agent in n8n takes each user request and decides which connector workflow to call, such as BigQuery, Snowflake or Google Analytics. About 90% of each connector workflow is a reusable template.
Actual tools: n8n, Bubble, AWS S3
Result as claimed: Adding a new data connector went from about a week to around a day and a half. David estimates 20 to 30 hours saved per month, and hundreds of hours overall, by building workflows in n8n instead of maintaining custom Lambda functions and API code.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Malaysia
He runs two Proxmox hosts, a NAS, a media server and five internet servers for his websites. Claude Code runs on its own small virtual machine and pulls jobs from his home dashboard, which offers 15 ready-made, read-only checks such as disk space, failed services, waiting updates and "What should I do first this week?". Each check runs now or on a schedule, and every job must finish before his 07:30 morning report.
Actual tools: Claude Code, Proxmox, SQLite
Result as claimed: In the first few hours it reported a monitoring service that had quietly failed on one internet server, listed security updates waiting on three machines in a sensible order, and pointed out that his main Proxmox host had no backup job at all, so he set one up. It also raised one false alarm, which he traced to his own fragile check.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Uruguay
Grok Bot is now his main inbox. It filters junk mail, tracks recurring costs, publishes each podcast episode to Substack, YouTube and Libsyn, helps book guests, and pitched podcasts he had been on about his book.
Actual tools: Grok Bot, Substack, YouTube, Libsyn
Result as claimed: About $4,600 a month cut across 75 subscriptions and services. Podcast publishing to Substack, YouTube and Libsyn now takes about 30 minutes, down from 2 hours, with the bot doing about 95%. It pitched 13 podcasts he had been on about his book: 5 replied and 4 said yes or welcomed the guest he suggested.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
A two-person team running four brands on five sales channels uses Claude and Grok to produce content across the brands, and built its own customer app in-house with Claude Code.
Actual tools: Claude, Grok, Claude Code
Result as claimed: More than S$14,000 a year in labour saved, up to 4x content output, about 60 manhours a month freed, and their own customer app built in-house, as stated in the award citation. The Straits Times (13 Nov 2026) reports that the 60 hours freed each month from repetitive tasks add up to more than $14,000 a year in labour savings, and that automated emails to shoppers who abandoned their carts or have not visited lately bring in an estimated $1,000 to $3,000 more each month.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
An operations co-pilot built on n8n and Claude extracts supplier invoices and checks them against master prices, runs finance workflows, keeps a multi-channel order dashboard and drafts social posts checked against a policy.
Actual tools: n8n, Claude
Result as claimed: 83% less time on invoices, about 40 manhours a week saved and 85% return on investment in the first year, as stated in the award citation.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Italy
Every day at 8am the bot reads his team's code commits and sends him a summary for each client, written for the person who will read it. He checks it and copies it into the client report.
Actual tools: Grok Bot
Result as claimed: Writing daily client reports took about half an hour. Now he spends about 5 minutes checking and copying the bot's 8am summary of the team's commits. He puts it at 20+ minutes a day, over 7 hours a month.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
A WhatsApp agent called The Nami Concierge answers enquiries and handles reminders and payment follow-ups that two admin staff used to handle.
Actual tools: WhatsApp, AI agent (model not named)
Result as claimed: Up to 83% less admin time, 50% fewer no-shows and 100% on-time payment, against a starting point of two admin staff spending about 60 hours a week, as stated in the award citation.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Restoke.ai reads supplier invoices, records stocktakes on mobile across outlets and updates each dish's cost with live supplier prices, so menu margins are visible as prices change.
Actual tools: Restoke.ai, StaffAny
Result as claimed: 100% return on investment within five months, 80% faster invoice processing and 40% less time on stocktaking, as stated in the award citation.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
An order automation called Milo handles personalised gift orders coming in from Shopee, Lazada and Shopify, about 2,500 a month.
Actual tools: Shopee, Lazada, Shopify, AI agent called Milo (model not named)
Result as claimed: 80% less rework, more than 20 hours a week saved and up to $5,800 a month saved, on about 2,500 orders a month across Shopee, Lazada and Shopify, as stated in the award citation.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
MeetGeek records client meetings and Claude turns them into handover documents. The agency also uses Higgsfield, AdCreative.ai and Blip AI for video and ad production.
Actual tools: MeetGeek, Claude, Higgsfield, AdCreative.ai, Blip AI
Result as claimed: A handover document that took 3 to 4 hours to write now comes from the meeting recording. The citation reports 85% faster onboarding, about 85% less video production time and up to $10,000 a month saved.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United States
Claude Cowork took control of his Mac and, with iPhone Mirroring, his phone. It audited disk space, Chrome and phone security settings, explained each issue and fixed them one at a time after he said yes.
Actual tools: Claude Cowork, iPhone Mirroring
Result as claimed: Cleared 125GB off a full drive, including a 104GB old work folder and a 14GB stale Google Drive cache. A 20-minute iPhone audit found nine settings already right and six to fix, including an uninstalled security update and Advanced Data Protection off. It refused to type his passcode or save his recovery key in email.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United States
He built his own campground reservation system, Keepr, with AI, starting with ChatGPT and moving to command-line agents and Claude Opus 5.5, and runs his campgrounds on it.
Actual tools: ChatGPT, Claude Opus 5.5, Vercel, Railway, Stripe Connect
Result as claimed: He built the Keepr reservation system with AI, starting with ChatGPT and moving to command-line agents and Claude Opus 5.5, and runs his campgrounds on it. Costs he lists: Vercel $537 in April and Railway about $125 a month. He found Stripe Connect fees he had missed and changed his fee structure.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
Eight Grok Bots found and scored roles, watched company career pages for new posts, mapped warm intros, drafted applications and cover letter points, researched interviewers and filed call notes. None was allowed to send anything.
Actual tools: Grok Bot
Result as claimed: Over a month: 90+ open roles found and scored, 70 people mapped for possible warm intros, 20 full application packages drafted with cover letter content, and notes filed on 35 calls. He puts the manual effort at about 80 hours. He wrote and sent every email and application himself.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Southeast Asia
Grok Bot analyses the top 10 articles for a keyword, drafts an article from his own notes, edits it with him live on a 30 to 60 minute call, and adds his own photos.
Actual tools: Grok Bot
Result as claimed: From blank document to a personalised article in about 3 to 4 hours, down from 1 to 2 days before.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
On top of his own Python pipeline, Claude Code loads bank statement files from several Singapore banks into YNAB, sorts transactions and asks him about unclear ones, learns new rules from his answers, and tracks each card's bonus cap per billing cycle.
Actual tools: Claude Code, YNAB, Python
Result as claimed: Monthly work across his Singapore bank and card accounts (downloading statements, importing to YNAB, sorting 200+ transactions, checking card bonus caps) used to take 2 to 3 hours and now takes 4 minutes. A 45-minute manual import became a 30-second command. His categoriser matched 40% of transactions on its own at first and 85%+ after a month of his corrections.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United States
He highlights the items he wants in his wholesaler's catalogue, photographs the pages and sends them to Muse, which builds the weekly ad. He does the same with the staff schedule he rewrites every two weeks and with employee time cards for payroll. Muse also writes some of the store's Facebook posts and posts to Instagram.
Actual tools: Muse for Small Business (Meta)
Result as claimed: The weekly ad used to take a couple of hours of typing and now takes about 15 minutes. In Meta's launch announcement he said photographing time cards and letting Muse add them up cut "at least 15 minutes from payroll."
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
The team uses Claude, ChatGPT and OpenMax to write first drafts of proposals, translate and localise content the same day (for example English and Chinese for Xiaohongshu), run checklists and sensitive-word checks before content goes out, and handle routine work such as prospecting, morning briefings and reports. People keep strategy, creative work and final checks.
Actual tools: Claude, ChatGPT, OpenMax
Result as claimed: From IMDA's SME AI Impact Awards 2026 infographic, as quoted on the agency's page: 2x higher employee productivity, 75% faster customer response times and 80% faster proposal creation. Proposal turnaround went from a full day to a few hours.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United States
Every hour Claude checks his Gmail for client requests, logs each one as a task in Clio and starts the work, for example drafting an NDA, a first-pass contract review or track-changes markup. By the time he opens a request, roughly half of it is done. Other scans swap Meet links for Zoom, block prep time before client work, send regulatory alerts and keep a privacy-sensitive client's names out of billing notes. Zapier MCP connects Claude to Clio, Gmail, Calendar, Zoom, Docs, Drive, Sheets, Xero and GitHub.
Actual tools: Claude, Zapier MCP, Clio, Gmail, Google Calendar, Zoom
Result as claimed: He estimates he does the same work in 20% of the time, about 5x faster, and Zapier's post says a task that took about five hours now takes one. Zapier counted 12,322 MCP tasks on his account since 3 April 2026, active on 100 of roughly 104 days. He keeps a 30-hour billable week and still bills by the hour, so he says clients get the shorter invoices.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
Scot and Susan Westwater run Pragmatic, an AI strategy consultancy with two full-time staff. Their first Lindy agent took under 20 minutes to set up and came back with about 25 conferences to speak at. Other agents draft and format proposals, edit client deliverables, compare technical options and write case studies.
Actual tools: Lindy
Result as claimed: As reported by Pragmatic: 20 to 30 hours a week saved, proposal turnaround down from a week and a half to under two days, conference research down from two weeks to 20 minutes, and a $50,000 project secured with Lindy's help. No staff were added.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Netherlands
Jonas runs three Claude Code workspaces for content, his AI agency and research, plus Make.com automations for a voice second brain, Substack tagging and a daily AI news roundup. In the week he tracked, agents helped pre-screen thousands of companies, research hundreds of them and draft personalised company reports for outreach, while he kept final emails and strategy in his own hands.
Actual tools: Claude Code, Make, Perplexity, Apollo, Obsidian, Figma
Result as claimed: In one tracked week he worked 28 hours while his agents put in about 90. He says he got about 4.2x as much done as in 2023. Agents pre-screened 6,000 companies and did deeper research on 250, and he says the setup can create 250 personalised company reports in two minutes. He also names lead generation as about 20x faster, research 4x, AI-assisted coding 8x and content 3x versus his 2023 baseline.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Leon’s studio built an AI Worker for a maritime services client. It watches the inbox for job specs, extracts line items with Claude, matches them to a rate card in Supabase, drafts a quotation, flags unsure items and waits for human approval.
Actual tools: Claude, LangChain, Supabase, Next.js, Vercel, Groq, OpenAI
Result as claimed: For a maritime services client, quoting vessel repairs used to take 2 to 4 hours while an ops manager checked 200+ line items against a rate card in Excel. The AI Worker Connective Labs built watches the inbox, extracts line items with Claude, matches them to the rate card in Supabase, drafts a quotation, flags uncertain items, and sends the draft for one-click approval. Leon says the draft cycle is 8 minutes including human review.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United States
Custom GPTs rank priorities, review the same document in a roundtable, and do preliminary legal drafting.
Actual tools: ChatGPT custom GPTs
Result as claimed: About 20 hours per week saved, called conservative by him. Training each GPT took about two weeks.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
An internally developed fixed RPA workflow retrieves the password, extracts and converts files, populates fields, matches accounting references and consolidates the reporting spreadsheet.
Actual tools: SMU internal RPA solution, Microsoft Excel
Result as claimed: GORU reports 60 min to 8 sec of data processing per run and 59 min saved per day.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Malaysia
Filum queries live order records, answers routine questions and prepares ticket tags and summaries.
Actual tools: Filum.ai AI Agent, Filum.ai Voice of Customer
Result as claimed: Filum reports 55%+ fully resolved without human handoff in Malaysia and Singapore in the first two months, and ~87% goal achievement to the team's criteria.
Read the story · Original source, opens in a new tab
Used in this story
Singapore
Volty built Cake Inspiration a WhatsApp agent that answers from the shop site, routes catalog shoppers to product pages, hands custom orders to a human, and learns from staff corrections Kai Yang approves.
Actual tools: WhatsApp, Volty / Voltade agent
Result as claimed: Voltade's 2 Nov 2026 case study says Cake Inspiration was answering about 90 WhatsApp inquiries a day by hand with median reply time up to 9 hours. The WhatsApp agent now answers from the shop knowledge base, points catalog shoppers to product pages, and hands custom orders to a human designer. Key results shown: median reply time 9 hours to 5 seconds, and 7 staff corrections applied in week one with no engineer in the loop. Founder Kai Yang is quoted saying he can fix the agent himself by telling it what he disliked about a reply.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
During a one-day Klicco workshop Sabirah used Claude to run bank-statement analysis and start unloading solo backend admin and content summaries.
Actual tools: Claude
Result as claimed: Klicco's public case study for Sabirah of Laiyyin.co, a Singapore stay-at-home mum and solo online business owner, says she completed a full bank statement analysis in under 15 minutes using Claude during a 1 May 2026 Mastering Agentic AI workshop, a task that previously would have taken days. The page lists results of 15 min bank analysis (was days) and 3 to 6 hours saved per week after the workshop. Her quote on the page: she will unload most backend work to AI and create summaries within minutes.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
Doneyli built a two-tier chief of staff: cheap rules catch urgent mail every 30 minutes, and a daily Claude pass classifies the rest and drafts replies with graduated trust.
Actual tools: Claude, Claude Code, Google Workspace
Result as claimed: In his free 15 Mar 2026 Substack post he says a two-tier email path (rules every 30 minutes, LLM triage once daily) cut LLM costs by roughly 80%. He reports under $3/day LLM cost for both tenants combined, about 50 emails/day per tenant, Tier 1 alerts in under 10 seconds, draft accuracy above 80% send rate without edits for graduated categories, and a Claude Max plan at $100/month covering 26 agents.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Mexico
Answered FAQs at any hour, suggested other products, sent photos and followed up when customers went quiet.
Actual tools: Meta Business Agent, WhatsApp Business App
Result as claimed: Guevara estimates an 80% reduction in time spent personally replying to customers.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
AI extracted WhatsApp orders, synced them into the existing platform and generated summaries for confirmation.
Actual tools: WhatsApp, AI-enabled Automated Ordering Management (product unnamed), Integrated accounting, inventory and sales platform (product unnamed)
Result as claimed: IMDA reports a 20% improvement in order accuracy and up to eight hours of labour saved per mistake avoided.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
The source describes document conversion and an accounting evidence PDF build. Public illustrations are invented-company mock-ups. Claimed private client inputs are not available for review. No completed client tax work or audit outcome is established.
Actual tools: Claude Code, PyMuPDF, Microsoft Excel
Result as claimed: Workflow demonstration, not an audited client result. The article illustrates the process with mock-ups whose company and figures are invented. Its claimed underlying client documents are not public. The source reports a first version of 91 pages with 15 accounts linked and a second of 164 pages with 20 accounts across 13 sections plus a five-document folio. It says trade-debtor and trade-creditor balances were flagged because they were absent from their reconciliation reports. The linked social post claims almost an hour saved, without independent timing evidence.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
He had a timer website with voice prompts and TalkBack support. A mobile browser could delay timers when the screen locked, so he asked an agent for a native app. He tested each build with TalkBack, reported what felt wrong, and the agent rebuilt it with two tabs, a foreground service, a wake lock and one persistent notification.
Actual tools: AI agent (not named in the post), Android TalkBack
Result as claimed: He says the agent turned the site into a native Android app in 2 to 3 hours, not a WebView wrapper. The first working build had meat and breathing timers, system notifications and exact alarms, notification permission handling, saved settings, TalkBack accessible controls and a test APK. He says he is not an experienced Android developer.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
Claude scans the media inbox in Gmail for relevant requests, researches trending community topics and judges what matters. Through Zapier MCP it saves reports to Google Drive, updates Sheets and posts Slack alerts.
Actual tools: Claude, Zapier MCP, Gmail, Google Drive, Google Sheets, Slack
Result as claimed: Zapier reports media inbox triage went from about 4 hours of human review each week to about 15 minutes.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United States
Claude, connected to Zapier through Zapier MCP, writes and publishes Zapier workflows from Ross's descriptions. The running workflows are scheduled Zaps that pull ServiceTitan data, flag bookings pending 20 minutes, send a 6:30 a.m. weekday brief and send matched calls and job revenue to Google Ads.
Actual tools: Claude, Zapier MCP, Zapier Next-Gen Zaps, ServiceTitan, Google Ads, Claude in Chrome
Result as claimed: Zapier reports 36 live Next-Gen Zaps running about 1,400 times a week across about 30 connected apps. Revenue went from about $300,000 a year to a projected $3.8 million in 2026, but neither source attributes that growth to AI.
Read the story · Original source, opens in a new tab
Used in this story
Unknown
After she uploads a video and TurboScribe makes a transcript, the agent suggests titles, drafts a short description, lightly cleans the transcript, opens her site, duplicates the template and fills in the video, title, description and transcript.
Actual tools: YouTube, TurboScribe, An AI agent the author does not name, run on her Claude subscription, Her website template
Result as claimed: She says it saves about an hour every week. She estimates an assistant for the same work would cost around €300 per month, against a Claude subscription of roughly $20 per month.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United Kingdom
Working from one folder of plain text files, Claude Code or Codex updates a note for each company, person and opportunity after every synced call transcript, creates tasks with priority and an estimated duration, starts promised work such as intro emails, and writes a daily brief that also arrives on Telegram. For webinars it makes the banners, drafts the descriptions and publishes the event to Circle and three Luma calendars once he says publish.
Actual tools: Claude Code, Codex, Obsidian, Granola, Wispr Flow, Telegram, Gmail, Circle, Luma, Lovable
Result as claimed: His own estimates, shared with Create With: proposal prep down from about 4 hours to about 30 minutes, an event banner and descriptions from about 40 minutes to about 5, and delivery effort down about 45% on one engagement. Average monthly revenue in the first half of 2026 ran at 2.4 times the 2025 monthly average, which he credits partly to a young business growing. His community is on course for about 33 virtual events in 2026, against 8 in 2025. The setup costs around £250 a month.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Odello reads customer orders sent on WhatsApp, uses the customer's history and the catering menu, asks clarifying questions, prices and confirms standard orders and enters them into the order system. It also helps staff draft non-standard replies and tracks delivery ETAs and proof of delivery.
Actual tools: Odello (eAI by TreeDots), WhatsApp
Result as claimed: eAI reports that two months after go-live around 92% of incoming orders are handled end to end by AI, order admin time is down about 85%, and around 30% of orders now arrive between 9pm and 9am.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Manus built a prep calculator that turns the weekend's orders into how much meat and other ingredients to buy, then grew it into a system he calls FLAME. It imports website and WhatsApp orders, flags addresses with no unit number, groups deliveries by area for freelance drivers and runs a WhatsApp inbox for confirmations and feedback.
Actual tools: Manus, Twilio WhatsApp Business API
Result as claimed: Manus reports that Mandai Smoke Co processes around 300 orders a month without hiring an admin team, and that the system was ready when the December holiday rush arrived. Chan says his drivers' hourly earnings are up to 90 SGD an hour because they follow the planned routes.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Philippines
For a Malaysian direct-to-consumer brand he replaced a chat tool with five self-hosted n8n flows across WhatsApp, Messenger, Instagram and Facebook. Rules route each question and Claude handles the unclear ones. An Apify and n8n pipeline moved customer reviews into the Shopify store.
Actual tools: n8n, Claude, Apify, Shopify
Result as claimed: He says about 70% of the brand's support questions are now handled without a person, five flows are live across the four channels, and 5,437 reviews moved to Shopify for $1 in API costs against $1,500 or more for a migration service.
Read the story · Original source, opens in a new tab
Used in this story
Unknown
Several times a day Claude picks active jobs from the firm's applicant tracking system, reads the company and role success profiles and segment notes, and uses the firm's sourcing tools through Zapier MCP to build a candidate list. A second Claude run checks each list later in the day. A separate process on the firm's own server reads recruiter notes and the Zoom transcript, scores the candidate, researches past employers and writes the submission. A Next-Gen Zap runs at 2 a.m. on weekdays to build prep sheets for screening calls and post a Slack note.
Actual tools: Claude, Zapier MCP, Zapier Next-Gen Zaps, Slack, Zoom, Google Sheets
Result as claimed: Curtis Key says human sourcers' candidates are approved 78 to 83 percent of the time and Claude's about 94 percent. He estimates the sourcing and quality-control automations save the team 240+ hours a week at current volume, and the submission process saves each recruiter 4+ more hours a week. Zapier reports sourcing output up about four times.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
Sheila reads contractor invoices from email, records them in a Google spreadsheet, sends payments in fiat through Mercury and in Bitcoin through Kraken, Lightning and Boltz, generates invoice PDFs, archives everything to Google Drive, submits expenses to OpenCollective, tracks P&L, generates 1099 reports and emails contractors when they are paid. It chains over 50 small scripts using a 600-line AGENTS.md that describes the workflows in plain English.
Actual tools: OpenCode, Mercury, Kraken, Google Sheets, Google Drive, OpenCollective
Result as claimed: Gleason says Sheila replaced the entire accounting flow at Soapbox and calls it a drastic time-saver. He gives no hours or dollar figures.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Before each writing session Claude sends her three ideas from her idea bank. After her email goes out it prepares a Substack draft and Facebook and Instagram versions in her own words. On Sundays it checks what actually went out and tidies the idea bank.
Actual tools: Claude
Result as claimed: Faith Teo says going from idea to posts on three platforms used to take five to six hours and now takes one to two, depending on how much polishing she wants.
Read the story · Original source, opens in a new tab
Used in this story
Singapore
Each night Claude gathers the WhatsApp messages she starred and files them into her to-do app under a Promises label, tagged by client. It also looks through her client chats and tells her who has gone quiet.
Actual tools: Claude
Result as claimed: Faith Teo says the planner page she used to check several times a day is retired and her head feels lighter. No time or money figure is given.
Read the story · Original source, opens in a new tab
Used in this story
Singapore
Every night Claude sorts her email into folders such as receipts, finance and a Wins folder for new subscribers. When she corrects a decision it remembers.
Actual tools: Claude
Result as claimed: Faith Teo says she used to check her email over and over and now clears everything in one sitting, once a week. No time figure is given.
Read the story · Original source, opens in a new tab
Used in this story
Spain
Claude Code skills sort his inbox and draft replies in Gmail, book supplier invoices into his invoicing software, check two bank accounts three times a day and flag payments it cannot match.
Actual tools: Claude Code, Gmail, contix
Result as claimed: Martin Schenk says invoices are booked and both bank accounts checked before his first coffee. He says the server service that reads invoices and bank accounts used small models through the API for 5,193 runs in 90 days, costing 2.50 dollars in total. He switched off direct sending in October so every email is now a draft.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
Claude routines run on set days to post content, upload expenses to accounting software, prepare and follow up meetings and handle her work email three times a day with replies ready to review.
Actual tools: Claude, n8n, Wispr Flow, Notion AI, NotebookLM
Result as claimed: Claudia Quintela says the manuals she wrote in January are all archived because none of that work is done by a person anymore. She says SEO work Claude now does brought her many new leads with a direct impact on the business in money, and that Claude sorted hundreds of thousands of family photos in a weekend. In March she wrote that AI was saving her thirty hours a week. No new figure is given.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Agents cut raw footage into captioned reels, schedule them to six platforms, collect views and comments, turn call notes into CRM follow-ups and send lead reports. A Telegram assistant logs spending and runs checks.
Actual tools: Claude Code, Hermes, Telegram, Zernio, Attio, Google Sheets
Result as claimed: Keith Loh says he runs seven agents on 56 skills in his workspace as of 28 October 2026, and that 115 reels have a scheduling record on his disk. His site claims 500K+ views on his channels in under 3 months of posting and 200+ posts a month across 7 accounts.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
Every three hours a bot checks his sites for new articles, writes copy per platform and queues it for publishing through Zernio. Another reads and triages two Google Workspace inboxes through the browser, including the spam folder, and drafts replies. Another builds invoices from forwarded client emails using PayPal and Wise. A fourth pulls client numbers through the Semrush MCP server or dashboards and drafts report commentary.
Actual tools: Grok Bot, Zernio, Gmail (Google Workspace), PayPal, Wise, Semrush MCP
Result as claimed: Fakhri says the four workflows give back 35 hours a week. He puts the social workflow at about 5 hours a week and the invoicing one at "the most annoying hour of my week". He says the bot found 48 real messages in spam in the first month, including a network approval he had chased for two weeks. He says it cost twenty dollars to start.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Malaysia
When someone books a demo, an agent researches the shop and briefs the caller. Other agents rank every open lead into a live call list that re-sorts about every 30 seconds, so a promised callback rises to the top. After each call the conversation is transcribed and the lead record is filled in. He also replaced a couple of hundred Zapier automations with his own stack in about four weeks.
Actual tools: Claude Code, Salesforce
Result as claimed: Fong says the average time to close out a lead is now about 1 minute, down from about 15. He says reps make two times more calls a day and meetings booked per 100 leads went up about 25%, which he says shows up in the bottom line. He says the marketing team used to spend 30 to 40% of its time fixing broken Zaps.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
He sends a screenshot, PDF or forwarded message with a short request. The agents have access to his Gmail, WhatsApp and Drive and can use a browser. Examples he lists: tracing an unexplained bank seizure to a tax charge, checking a tax portal and posting a document by registered mail, reviewing and approving year-end company accounts, and flagging terms in a merger engagement letter that contradicted what was agreed by email.
Actual tools: Claude Code, Codex, Gmail, WhatsApp, Google Drive
Result as claimed: He estimates the AIs saved him at least 2 days of work over the past month. He says the error rate seems similar to or lower than his own and that agents rival a human PA for many tasks.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Netherlands
Claude Code reads a few plain notes about his business, clients, sales, marketing and finance before each task. It updates his Trello board, briefs him before each meeting with who, why and where they left off, labels email by intent and drafts replies in his voice, builds n8n workflows from a plain description, and replies to and qualifies leads.
Actual tools: Claude Code, n8n, Trello, Obsidian
Result as claimed: He says a personalised cold email went from 30 to 45 minutes to 2 minutes, a Monday LinkedIn post from 2 hours to 15 minutes and client research from 1 hour to 5 minutes. He says proposals get drafted while he makes coffee.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United States
ChatGPT surfaced cross-business context, briefed priorities and helped with email and financial questions. Claude drafted proposals from notes and scope.
Actual tools: ChatGPT, Claude, QuickBooks
Result as claimed: Koehler estimates AI saved him 60 to 75 hours in August 2026. He says this is his own rough reconstruction of the month, not a time log.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
The chatbot answered routine website enquiries from a staff-maintained knowledge base.
Actual tools: Unnamed generative AI website chatbot
Result as claimed: CNA Brand Studio reports response times fell from about 30 minutes to under five minutes, chat traffic rose about 60 per cent and appointment bookings rose roughly 20 per cent.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
A custom system consolidated and deduplicated records, populated profiles, supported segmentation and chatbot responses. Fixed automation sent reservation confirmations.
Actual tools: Unnamed AI-enabled CRM, WhatsApp
Result as claimed: Mister Mobile estimates it saves about 120 man-hours of manual work a day by automating data consolidation, customer replies and marketing broadcasts, after merging more than 585,000 customer records and 1.05 million orders.
Read the story · Original source, opens in a new tab
Used in this story
Unknown
Claude, connected to approved client books through NumbersGame, handles transaction review, month-end entries for prepaids and depreciation, custom reports, KPI review and board decks. He turned his process into a Claude skill that runs on the third day of each month.
Actual tools: Claude, NumbersGame, QuickBooks Online, Xero
Result as claimed: By his own estimate, his day-to-day accounting through month-end reporting dropped from roughly 10 to 15 hours a month to roughly one to two hours a month in his solo-firm workflow. He says a 12-month report that drills from ledger line to vendor to transaction is now a regular client deliverable.
Read the story · Original source, opens in a new tab
Used in this story
United States
An AI receptionist she built answers calls, books estimates on a calendar and forwards upset callers to her. A freelancer built AI customer relationship software for the business.
Actual tools: ChatGPT, AI receptionist (tool not named in source)
Result as claimed: Inc. reports she says the automations saved her 20 hours a month. It reports the couple rebuilt revenue to $60,000 to $65,000 a month after it fell from $70,000 to $40,000, while working 80-hour weeks before the automations.
Read the story · Original source, opens in a new tab
Used in this story
Unknown
Each Tuesday he runs a Claude skill on the newest A/R aging export. It keeps a running invoice ledger, checks his email for follow-ups and replies, asks him only about missing pay dates, drafts one follow-up per customer after 7 days without a reply and builds a week-by-week cash forecast. It once suggested asking a general contractor for early release of retention.
Actual tools: Claude, Claude project he calls CEOAi, Email (checked for replies)
Result as claimed: He says average days for open receivables are down more than 11% and weekly receivables work fell from 1 to 2 hours to 10 to 15 minutes. One early retention release email the skill suggested brought a check within 30 days.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
United States
AI tools taught her startup basics, helped her write a business plan and helped her refine a pitch for early-stage investors.
Actual tools: AI tools (not named in the source)
Result as claimed: Reuters reports that funding began to flow and that Here Now Health, launched in January 2025, had 16 employees and certification in three states by July 2026. Reuters gives no funding amount and does not separate the part AI played. She called it "like going to a master's level class every day from the robot."
Read the story · Original source, opens in a new tab
Used in this story
Singapore
The custom platform connected business data. Its AI modules allocated inventory and coordinated scheduling, quotes, invoices and vendor work. The showcase also attributes gains to broader automation.
Actual tools: AutomateAI (custom platform built with Tridorian), Google Vertex AI, BigQuery, GenAI web enquiry chatbot
Result as claimed: IMDA's 2025 Digital Leaders Showcase reports a 67% cut in order processing time, 850 staff hours saved each month through automation, a 65% rise in customer retention and a 257% return in the first year. An earlier enquiry chatbot saved 82% of the time spent on manual replies.
Read the story · Original source, opens in a new tab
Used in this story
United Kingdom
Transcribed voice notes, drafted channel-specific posts and prepared branded carousels. A scheduler published approved posts.
Actual tools: Claude, OpenAI speech-to-text, Cloudflare, Supabase, Instagram API, LinkedIn API
Result as claimed: The agency that built it reports 30 drafts written from her voice notes and 11 posts published, counted from the live system on 12 November 2026, three days from approved plan to a live app. She says the pilot lets her save time making content, without a number.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
France
Sent a morning brief, checked connected apps every few hours, suggested and ran tasks in channels, kept artifacts such as his to-do list and opportunity dashboard updated, and flagged a missed client request to add a meeting participant.
Actual tools: Claude Duet, Claude Cowork, Gmail, Google Calendar, Claude in Chrome, Google Drive
Result as claimed: He reports that Duet estimated about 40 hours saved in three weeks and counted about 5 hours lost to rework. He calls it roughly one week of work he would not have had time to do.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Answered straightforward customer enquiries through a GenAI chatbot.
Actual tools: Websentials Omnibot
Result as claimed: IMDA reports weekly hours on customer enquiries fell from 18 to 8, staff handling queries went from three to two, and online sales rose 80%.
Read the story · Original source, opens in a new tab
Used in this story
Singapore
AI helps prepare personalised communication to parents. The centre also uses AI across curriculum development, marketing, customer service, HR and operations.
Actual tools: AI tools, names not stated
Result as claimed: IMDA's award citation says the centre cut personalised parent communication preparation time by over 80%.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Ready-to-use AI tools support image and video production and faster iteration inside the agency's creative production.
Actual tools: Ready-to-use AI tools, names not stated
Result as claimed: IMDA's award citation says campaign production time fell from three to four weeks to three to seven days, while output rose by over 200%.
Read the story · Original source, opens in a new tab
Used in this story
United States
Audited her email automations, shut down duplicate flows that had been emailing customers twice, rebuilt her templates and turned on tracking. She also says it rebuilt her website in about half an hour and expanded her two-page brand guide to 12 pages.
Actual tools: Muse for Small Business (Meta)
Result as claimed: Duplicate email flows that had been double-emailing customers were shut down before the holiday rush. She says the website rebuild took about half an hour. Earlier she says she spent thousands of dollars on contractors whose work was not done correctly.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Vietnam
Estimated storage needs, checked availability, replied in four languages, followed up incomplete bookings and updated booking records.
Actual tools: STOW
Result as claimed: MyStorage reports that in Q2 2026 STOW cut first response time from 3 minutes 40 seconds to 42 seconds, wrote 56% of outgoing replies and saved about 89 hours of manual work a month. It says conversion reached about 7% against about 4% for fully manual chats the year before.
Read the story · Original source, opens in a new tab
Used in this story
United States
AI by Zapier reads each new lead email, identifies the case type and drafts a response with qualifying questions. Fixed automation creates the Clio Grow record, sends a text and email, calls the lead and alerts intake staff.
Actual tools: Zapier, AI by Zapier, Clio Grow, RingCentral, Gmail
Result as claimed: About two minutes from form submission to outreach. COO estimates more than 66 admin hours a month saved, based on a 99.46% successful run rate. Conversion rose from 11.06% in September to 13.58% in May. More than 80% of calls answered within 20 seconds in 15 or more of the past 20 weeks.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Singapore
Answers routine customer enquiries right away, including when the founders are teaching or away.
Actual tools: Meta Business Agent
Result as claimed: Routine questions that took around an hour or longer now get almost immediate replies. The founders estimate 10 to 15 working hours saved each week.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Japan
Built a re-runnable pipeline that pulls the month's transactions over MCP, matches bank rows to receipts by amount and date, proposes a category with a one-line reason for each row and flags rows it isn't sure about.
Actual tools: Claude Code, MCP connection to his accounting platform (platform not named)
Result as claimed: 200 transactions reconciled and 11 miscategorised, a 94.5% hit rate. Review shrank to 18 flagged rows plus a sample. 7 of the 11 errors were in the flagged pile and 4 were caught on the sample.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
Scheduled agents triage issues, review search console and analytics for anomalies, write and publish blog posts, run SEO audits, flag Google Ads campaigns and post growth content. They propose tasks into a backlog.
Actual tools: Claude (scheduled agent system he built)
Result as claimed: About 60 hours a week before, about 15 to 20 hours after one month. Agents handle roughly 70% of his manual work. Organic impressions roughly doubled in a month. He shipped three features and two new MVPs in the month.
Read the story · Original source, opens in a new tab
Used in this story
Singapore
Once a week it reads his spam and archived email, groups it by sender and looks for events and articles that fit his goals for the year. For each useful sender it gives up to 5 action points and suggests pulling it out of spam or unsubscribing. It now checks his calendar and working hours before it shortlists events. Claude also names new files and chats to one convention and drafts unusual workshop formats for him to pick from.
Actual tools: Claude
Result as claimed: The first run went through 718 emails and it now handles about 30 a week. He says AI gives him back 3 to 4 hours a week, and one webinar it flagged put him on a path to new work.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
Reads new email and routes each one. Training enquiries become a HubSpot contact and deal with a drafted reply. Community messages get a draft reply. Possible collaborations become a Todoist task. Invoices and admin get a label. Also runs a scheduled Monday AI news digest.
Actual tools: Claude, Gmail, HubSpot, Todoist, Notion, Riverside, Gamma
Result as claimed: She says agents save her between 3 and 4 hours a week.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
Each morning a headless Claude Code run takes the oldest screenshot, classifies it, writes a 45 to 115 word Note in his voice, checks it against copy gates, uploads the image, schedules it for the next free day, reads the schedule back and archives the screenshot. Screenshots with emails, phone numbers or orders are set aside.
Actual tools: Claude Code, SubflowAI Substack connector, macOS scheduler, Substack Notes
Result as claimed: 29 Notes scheduled by 23 Sep and 38 by 12 Oct with none written by him. 12 measured Notes had a median of 3.5 likes against 5 for his other Notes, which he reads as no lift.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Unknown
Claude, connected to the company's systems, pulls data for board meeting prep, the monthly close and weekly leadership dashboards and presents it. Custom apps built with Claude replaced Monday, ClickUp and Airtable.
Actual tools: Claude, Claude artifacts, Claude Tag (Slack), Jira, Amplitude, Metabase, SigNoz, GitHub, Iterable, Google Drive, Excel, PowerPoint
Result as claimed: For the last two quarters Claude has built a 70-slide board deck with one click, and the monthly close went from 2 weeks to 2 days.
Read the story · Original source, opens in a new tab
Used in this story
United States
The agent collected competitor ad examples and AI labelled creative assets. A separate fixed Make scenario created ad structures and uploaded approved assets.
Actual tools: Make, Make AI Agent, Gemini, Google Sheets, Meta Ads, TikTok, Google Ads, Apify, Slack
Result as claimed: Make reports a 5 to 7 times cut in time spent on Meta ad setup, and creative labelling of 1,000+ assets a year going from weeks to seconds. The 1,000+ figure refers to 2025 volume.
Read the story · Original source, opens in a new tab
Related skill
Used in this story
Norway
Claude queried normalised live marketing data via Supermetrics and assembled report analysis.
Actual tools: Claude Cowork, Supermetrics connector, Google Sheets
Result as claimed: Anthropic's case says client reports that took 10 hours of manual spreadsheet work now take 20 minutes, and 17 of those minutes go into dropping the finished data into the client's Google Sheet.
Read the story · Original source, opens in a new tab
Related skill