Share your best AI workflow. We could show it to 2M+ people.

I changed my iPhone’s Action Button from mute/unmute to a menu of shortcuts. When I press it, I can select a shortcut I created called “Expense AI.” Expense AI records my voice, appends the transcript to a prompt, and sends it to ChatGPT. ChatGPT returns a JSON object containing the time, amount, expense description, payment method, and category. The data is then saved to a live Google Sheet. Step-by-step: 1. I changed the iPhone Action Button to display a menu of shortcuts instead of controlling mute/unmute. 2. I selected the “Expense AI” shortcut from that menu. 3. Expense AI recorded my voice and appended the transcript to a prompt. 4. It sent the prompt to ChatGPT, which returned JSON with the time, amount, expense description, payment method, and category. 5. The resulting data was saved to my live Google Sheet.

Planning a trip often meant piecing together information from a dozen different places. I might save inspiration from Instagram Reels and carousels, bookmark useful links in my browser, receive booking confirmations as PDFs from different platforms, and have tickets spread across emails and travel apps. By the time the trip got closer, I had all the information I needed—but no easy way to see it together or turn it into a coherent plan. I recently started using an AI travel tool called Mindtrip to solve this. It’s free and brings my saved content, links, bookings, PDFs, and tickets into one place. Mindtrip uses AI to understand the context and organize everything into a single trip, including the tips and hacks in Instagram Reels. It then creates a day-by-day itinerary based on what I’ve saved, liked, and booked. I have one place to see, manage, and adjust my entire trip instead of constantly switching between different sources. It also provides reminders and checklists based on my plans, along with a map view of where I’m supposed to be going. Step-by-step: 1. I saved travel inspiration from Instagram Reels and carousels, bookmarked useful links, and collected booking confirmations, PDFs, and tickets from different platforms, emails, and travel apps. 2. I brought the saved content, links, bookings, PDFs, and tickets into Mindtrip. 3. Mindtrip used AI to understand the context and organize everything into a single trip, including tips and hacks from Instagram Reels. 4. I used the resulting day-by-day itinerary, based on what I had saved, liked, and booked. 5. I used Mindtrip’s reminders, checklists, and map view to manage and adjust the trip in one place.

I built a complete work order platform using Claude Cowork, ChatGPT, and Base44 for my walk-in laser engraving and UV/DTF printing company. The platform takes customers from their initial meeting through final pickup, captures all the required project information, and prints a work order that travels with the project throughout the shop. We use connectors for Twilio SMS messaging, Gmail email communications, Square payments, and Shopify to capture online sales. Customers receive an SMS message the first time they visit and are kept informed throughout the process. We use a Kanban-style board to move projects through each stage. We call the platform FlowTrax.app. It tracks all required customer information, each request, and all communications associated with every sale. It also brings our online sales into the same workflow so we can track their progress. The platform has made us more efficient, and we no longer lose previous customer information. Step-by-step: 1. I built the platform with Claude Cowork, ChatGPT, and Base44 for our laser engraving and UV/DTF printing company. 2. I capture the customer’s information, project request, and required project details during the initial meeting. 3. I print a work order containing the required information so it can travel with the project through the shop. 4. I use Twilio for SMS messaging and Gmail for email communications throughout the process. 5. I use Square to process payments and Shopify to bring online sales into the workflow. 6. I send customers an SMS message when they first come in and communicate with them throughout the process. 7. I move projects through the workflow using a Kanban-style board until final pickup. 8. I track customer information, requests, communications, and the progress of both in-person and online sales in FlowTrax.app.

I wanted to learn how to build websites and sell them, but I didn’t know where to start. I used AI—specifically DeepSeek—to help me plan a roadmap. Because I already had experience prompting large language models to get the results I wanted, I asked DeepSeek which areas of knowledge I would need for this path. I also asked it to prioritize each area using statistics and facts. I reviewed the areas I didn’t know and prioritized them, then told the AI that I needed free resources only. I asked it to rank the topics based on what I didn’t know or understood the least. Finally, I asked it to create a Markdown file with all the resources formatted as checklists and imported the file into Notion. Now I have a plan I’m following instead of a “someday I’ll do this, hopefully” idea. Step-by-step: 1. I explained to DeepSeek that I wanted to learn how to build and sell websites but didn’t know where to begin. 2. I asked it to identify the areas of knowledge I would need for that path. 3. I asked it to prioritize those areas using statistics and facts. 4. I reviewed the topics I knew the least about and used that information to prioritize them. 5. I specified that I wanted free learning resources only. 6. I asked DeepSeek to create a Markdown file listing the resources as checklists. 7. I imported the Markdown file into Notion and started following the resulting plan.

Several months ago, my dad was in and out of the hospital. My two brothers and I were trying to coordinate his doctor appointments, manage his medications and potential interactions, and keep track of all the other details involved in his care. At one point, up to seven different doctors were seeing him in the hospital on any given day. It became important to track every medication he was taking, what each one was for, and information such as his weight and other vital statistics. When he returned home, we also had to make sure someone checked on him and his wife every day, helped him stay on schedule with his medications, and recorded his diet, mood, and weight. We initially used Apple Notes, a shared iCalendar, multiple text threads, and a weekly call between the three of us. The mental load was huge. If we needed to find information from the previous week, we had to scroll through pages of Apple Notes to locate it. We also struggled to keep the rest of the family updated. Before my dad passed away, I started using Claude Code and Codex to build a simple tool that would keep everything organized and searchable. It also displayed trends in areas such as his mood, appetite, and vital signs, and included a calendar showing who was covering which days and times. The tool was still fairly basic when he passed away. Afterward, we encountered the administrative headaches involved in closing out his estate. It was far more complicated than we expected. We thought having a will, power of attorney, and other documents meant we were prepared, but we were wrong. I began integrating those lessons—and the things we learned not to do—into the final product, Eldercare Log: eldercarelog.com. I built the final tool with Claude helping draft a PRD, which I then handed to Codex for the coding work. It took a few weeks of refining the product with Codex. The tool is hosted on Vercel, with Supabase and Stripe on the backend, and includes the security features I built into it. It is the tool I wish had existed when my brothers and I were going through this journey before my dad’s passing. Step-by-step: 1. I coordinated my dad’s doctor appointments, medications, vital statistics, and other care details with my two brothers while he was in and out of the hospital. 2. We tracked his medications, their purposes, his weight, diet, mood, appetite, and other vital signs while he was at home. 3. We coordinated daily visits and coverage using Apple Notes, a shared iCalendar, text threads, and a weekly call. 4. I used Claude Code and Codex to start building a searchable tool that organized his care information and showed simple trends. 5. I added a calendar to track which family member was covering each day and time. 6. After my dad passed away, I incorporated what we learned from handling his estate, including the things we wished we had known earlier. 7. I used Claude to draft a PRD, then gave it to Codex to handle the coding work and refined the product with Codex over several weeks. 8. I built the final tool with Vercel, Supabase, and Stripe on the backend.

I love playing poker, both online and live. One month, I performed poorly. Although it felt like the cards were running badly, I wondered whether I had developed a problem in my game and was blaming my losses on bad luck. I asked around, including asking AI, whether a tool existed that could measure luck from poker hand histories. The unwelcome answer was that it did not. I'm not a coder, but after doing some research into vibe coding, I started building a luck-detection app in a Claude chat. Claude built the UI directly in the chat window and advised me on the formulas I was using to calculate luck for the cards dealt, my performance on the flop, and my performance when I went all in. All three metrics have strong averages, and luck is what varies them. I used a bell curve to model hand outcomes and a Monte Carlo simulator, which Claude suggested and executed, to evaluate all possible outcomes. The result astonished me because it was so useful. I immediately fixed two major leaks in my game and felt better knowing that bad luck really was the main problem affecting my results. I liked the tool so much that I decided to turn it into a full web app with Claude Code, and now an iPhone app that I may let other people use for free. I also had a lot of fun building it—except for learning how to use Xcode. That was a pain, even with step-by-step guidance from Claude. Step-by-step: 1. I reviewed a month of poor poker results and questioned whether bad luck or problems in my game were causing the losses. 2. I researched whether a tool existed that could measure luck from poker hand histories and learned that I would need to build one myself. 3. I used vibe coding to start building a luck-detection app in a Claude chat. 4. I had Claude create the UI and advise on formulas for evaluating cards dealt, flop performance, and all-in performance. 5. I used a bell curve to model hand outcomes and a Monte Carlo simulator to evaluate possible outcomes. 6. I used the results to identify and fix two major leaks in my game and confirm that bad luck was also affecting my results. 7. I expanded the project into a full web app with Claude Code and then began building an iPhone app, working through the added challenge of learning Xcode.

I built AIEnhancer to solve a problem I often encountered when working with low-quality images. Many images contain useful content but are too small, blurry, or lacking in detail to reuse effectively. The project uses AI-based image processing to improve image resolution and recover visual details. I wanted to make the workflow simple: upload an image, process it, and receive an enhanced version without needing professional image-editing software. One challenge was finding the right balance between sharpening details and avoiding artificial-looking results. During development, I experimented with different enhancement approaches and focused on keeping the output natural. I'm still interested in improving enhancement quality for different types of images. I'd like to hear how other developers handle image restoration and super-resolution, especially for difficult or heavily compressed images. Step-by-step: 1. I identified the problem of reusing images that were too small, blurry, or lacking in detail. 2. I built AIEnhancer to process low-quality images with AI-based image enhancement. 3. I designed the workflow around uploading an image, processing it, and receiving an enhanced version. 4. I experimented with different enhancement approaches to improve resolution and recover visual details. 5. I evaluated the results for a balance between sharper details and a natural appearance. 6. I continued exploring ways to improve enhancement quality for different image types, including difficult or heavily compressed images.

I turned an open-source XML lookup fix into a debugging walkthrough that other developers can run. I used Codex, Python, and GitHub to document a bug where a document style containing quotation marks could be saved but then fail during retrieval because its name was interpolated into an XPath expression. Step-by-step: 1. I gave Codex the original patch and inspected the affected code, tracing the stored value through the public API, the wrapper, and the XML library. 2. I extracted the smallest standalone reproduction with `lxml`. It included a name containing both single and double quotes and kept the failing lookup visible. 3. I replaced interpolation with a bound variable: `styles.xpath("style[@name=$name]", name=name)`. I checked that the wrapper forwarded variables while preserving its namespace mapping. 4. I ran both versions. The interpolated version raised `XPathEvalError` in my reproduction. Binding the value matched five exact names, while an absent name returned no match. 5. I published the explanation, executable example, and upstream patch together. I inspected the rendered article and links and disclosed AI assistance. A prompt to reuse: "Reduce this lookup failure to a runnable example. Keep the failing case, show the fix, and check ordinary text, both quote types, custom namespaces and a missing value. Report which checks actually ran." The original fix had already been merged, so this workflow documents it. Broader service behavior needs separate tests. I used my existing Codex and local Python setup and did not measure time savings. I am Hồ Khắc Huy, a freelance software engineer. This is my independent open-source work. The linked article includes the runnable example, upstream contribution, and my contact details: https://github.com/builtbyhuy/builtbyhuy/blob/main/notes/2026-09-06-xpath-variables.md

I've been thinking about the value of handoff documents and explanatory documents as we continue exploring efficient ways to work alongside AI to build software and improve communication. Even though we use many different tools, Markdown still has an important place. This new version of an HTML handoff document can document what exists, describe what could exist, or serve as a mockup for a brainstorm. It lets us communicate with remarkable fidelity through visuals, hierarchy, and formatting. It's also an efficient format for AI to understand. We shouldn't underestimate the significance of AI communicating with us through a visual medium. A visual flowchart with thoughtful design, layout, animation, and progressive disclosure can help us understand the logic and flow of incredibly complex systems much faster. I tried all kinds of tools, including React Flow and Mermaid. They're fun to experiment with, but Archify is a game changer for this use case. I can point it at any technology, repository, or brainstorm and work with it to build flowcharts with animations and clean, distinctive design. It's also completely free. https://tt-a1i.github.io/archify/# Step-by-step: 1. I identify whether I need to document what exists, explore what could exist, or mock up a brainstorm. 2. I use Markdown and an HTML handoff document to communicate the ideas with visuals, hierarchy, and formatting. 3. I consider tools such as React Flow and Mermaid for creating visual representations. 4. I point Archify at the relevant technology, repository, or brainstorm. 5. I work with Archify to develop a flowchart with animations, clean design, and progressive disclosure so the system's logic and flow are easier to understand.

I moved from public school to a private school with no autism-specific training, and I found students on the spectrum struggling with one-size-fits-all lessons. AI helped close that gap for both me and my students in about five minutes per lesson. I take a standard lesson and rebuild it around the individual student. I upload only the blank assignment—never any student data—and prompt AI to adapt it to that learner. I include their favorite colors and niche interests, replace abstract examples with personalized ones, add sentence starters, and allow them to draw responses instead of writing dense paragraphs. Step-by-step: 1. I take an existing lesson and create a blank version of the assignment. 2. I upload only the blank assignment to my AI tool, without including any student data. 3. I provide the student’s interests, preferred colors, and reading level. 4. I ask the AI to rebuild the lesson for that specific student. 5. I use personalized examples, sentence starters, and drawing-based response options to make the lesson more accessible. The result is no shutdowns, no walls, and a genuinely engaged and grateful student. One untrained teacher can now individualize any lesson for a unique learner in minutes.

A chocolate factory client had a simple but expensive sales problem: good conversations with buyers sometimes went quiet because nobody followed up at the right time. Some opportunities were waiting on sample feedback, pricing, private-label details, minimum order quantities (MOQs), packaging, or distributor discussions. The opportunities were still active, but they were easy to lose track of. I built a follow-up workflow for the client in Awish. Step-by-step: 1. I opened the Awish chat and wrote: “Track our open sales opportunities from Microsoft Teams and Google Sheets. Every business day, review the latest customer conversation, sales stage, planned follow-up date, customer interest, and how long we’ve been out of touch. Find opportunities that need follow-up around samples, quotations, private label, MOQ, packaging, or distribution. Explain why each customer should be contacted, recommend the next action, and prepare a personalized Teams follow-up message based on the real conversation history. Never send anything without the salesperson’s approval. When the customer replies, close the old follow-up action and update the Sales Master record.” 2. Awish understood the request, planned the workflow, and selected Microsoft Teams and Google Sheets for the process. 3. I connected the client’s accounts, reviewed the plan, and approved the automation. 4. When a new customer conversation takes place in Teams, Awish updates the related sales record. 5. Every business day, it reviews open opportunities and finds deals that are overdue, at risk, or ready for the next follow-up. 6. For each opportunity, it explains why follow-up is needed and what the salesperson should do next. 7. Awish prepares a personalized Teams message using the actual conversation history, but waits for the salesperson’s approval before sending anything. 8. When the customer responds, the previous follow-up is closed and the opportunity status in Google Sheets is updated automatically. Now the sales team gets a prioritized daily list of who needs attention, why they need attention, and what the next message should be—without manually reviewing every old conversation. I want to keep building more workflows like this for real businesses. If you have a complex, repetitive process in your company that you think should be automated but you’re not sure how to build it, send it to me. I’d be happy to see if I can turn it into a working automation with Awish.

AI is useful for researching investments, but most workflows stop at “What should I buy?” The harder part comes afterward: Does the idea make sense given what I already own? What would make me add to the position? When should I take profits? What evidence would prove the original thesis wrong? And how do I track all of that without constantly watching the market? I use AI to turn a one-time investment research session into an ongoing decision and monitoring system. I start by asking AI to research the market for potential opportunities. In my case, I look specifically for strong mean-reversion trades, but the same workflow could start with value opportunities, macro themes, sector rotations, individual stocks, crypto, or almost any other investment strategy. Then I give AI my actual portfolio so it can evaluate those ideas in context. After I decide which recommendations I agree with and manually make the trades, I have AI convert each investment thesis into explicit rules for what should happen next. Finally, I turn those rules into automated monitors that periodically check market conditions and alert me only when something happens that warrants another decision. Step-by-step: 1. I define what I’m looking for by asking AI to research potential investment opportunities using criteria I care about, such as mean reversion, valuation, momentum, macro conditions, risk/reward, or another strategy. 2. I have AI investigate current market conditions and rank the opportunities, narrowing a large universe down to a manageable set of ideas worth examining further. 3. I pressure-test each thesis by asking why the opportunity exists, what could drive the expected outcome, what the major risks are, and—most importantly—what evidence would invalidate the thesis. 4. I provide my current holdings so AI can identify overlapping exposures, concentration risks, hedges, or positions that conflict with the new ideas. 5. I ask AI which existing positions the research suggests reviewing and where new exposure might make sense. The goal is a small number of actionable decisions rather than a giant list of interesting trades. 6. I review the analysis and independently decide whether to buy, sell, hold, or do nothing. I keep actual trade execution under human control. 7. Before the market moves, I define the next decision for every position by asking AI to identify conditions that would warrant reviewing whether to: - Add - Take profits - Reduce exposure - Exit - Reconsider the original thesis 8. I turn those conditions into automated monitors. I have ChatGPT periodically check the relevant prices, yields, economic indicators, news, or other variables. Instead of sending routine updates, I tell it to alert me only when a predefined trigger occurs. 9. When a trigger fires, I return to the original thesis with the new information and decide what—if anything—should change. Instead of using AI for isolated investment recommendations, I now have a repeatable loop for managing an investment thesis over time: Find an opportunity → Understand it → Compare it to what I own → Make a decision → Define what would change my mind → Let AI watch for it The most useful part may actually come after the investment decision. By deciding in advance what evidence would make me add, take profits, or reconsider the thesis, I don’t have to start my analysis from scratch every time the market moves. AI becomes a persistent research and monitoring layer while I remain responsible for every investment decision and trade.

Design taste is hard to put into words. I may know immediately that one website, presentation, report, or graphic feels right and another doesn't, but still struggle to explain whether the difference comes from the typography, spacing, colors, density, layout, or overall aesthetic. That makes working with AI frustrating. If I can't describe what I want, the AI has to guess. Instead of trying to explain my design taste upfront, I turned the process into an iterative visual feedback loop. I first asked the AI to suggest existing websites with different design styles so I could identify examples I liked. Once it had those references, I asked it to create three substantially different HTML mockups for the same content. I picked the direction I liked best, explained what I liked and disliked, and had the AI generate another set of alternatives based on that feedback. After about three rounds, the AI had a much better understanding of my design preferences than I could have given it in a written prompt. The basic loop is: show me examples → I choose → generate alternatives → I react → refine → repeat. This is essentially preference elicitation through examples. Research on human-AI interaction has similarly found that people can refine difficult-to-articulate preferences by reacting to concrete alternatives rather than specifying everything upfront. Step-by-step: 1. I asked the AI for several visual references, including websites or designs that represented distinctly different styles. 2. I reviewed the examples and identified the ones I liked. I didn't need sophisticated design terminology; I simply described what I preferred and anything obvious that I liked or disliked. 3. I gave the AI something real to design. I used an actual report I was working on so I could evaluate the styles in context. 4. I asked the AI to create three substantially different visual directions rather than minor variations of the same design. 5. I chose the direction closest to my taste and explained what I liked, what I didn't, and which elements from the other versions I wanted to incorporate. 6. I asked the AI to generate three new alternatives using everything it had learned so far. 7. I repeated the process until the designs began to converge. I did roughly three rounds, with each round narrowing the design space and giving the AI more information about my preferences. 8. I asked the AI to summarize what it had learned into reusable design guidelines so future projects could start with those preferences. Instead of trying to translate an aesthetic preference into design terminology, I let the AI learn my taste from my choices. The important shift is simple: don't describe what you like—show, choose, react, and refine. This can work for websites, presentations, reports, branding, graphics, interior design, clothing, invitations, or almost anything else where I know what I like when I see it.

I built a custom AI knowledge-transfer GPT designed to prevent institutional knowledge from walking out the door when an experienced employee leaves. The trigger was an experienced operations manager preparing to leave the organization. I realized that while we had procedures, files, emails and account documentation, a huge amount of operational knowledge existed only in that manager’s head: customer preferences, recurring staffing problems, site-specific quirks, key relationships, historical issues, workarounds, lessons learned and the small details that can take a replacement months to discover. My goal was to capture that knowledge and turn it into an interactive resource for the incoming operations manager. First, I conducted and recorded an in-depth interview with the departing manager. Instead of only asking standard turnover questions, I had them walk through the operation as if they were personally training their replacement. We discussed customers, employees, locations, scheduling, recurring problems, escalation procedures, communication preferences, historical decisions and things they believed a new manager might not realize immediately. I then transcribed the conversation and used AI to analyze the interview for knowledge gaps. I asked the AI to approach the transcript from the perspective of someone taking over the job and identify important questions that had not yet been answered. Using those gaps, I had AI create customized knowledge-transfer questionnaires specifically for the departing manager. These asked more targeted questions such as: What problems happen repeatedly? What customer preferences are not documented? What exceptions exist to normal procedures? What mistakes is a new manager likely to make? What information exists only in your memory? What would you make sure your replacement understood during their first 30 days? The departing manager completed those documents, giving me another layer of institutional knowledge that would normally be lost. Next, I organized the interview transcript, completed questionnaires, operational procedures, account information, contacts, historical notes and other relevant documents into a knowledge base. I uploaded that information into a custom GPT and instructed it to function as an operational knowledge-transfer assistant. The GPT was told to base answers on the captured information, not invent answers when information was missing, and clearly tell the user when something needed to be verified. The incoming operations manager can now interact with that knowledge conversationally. Instead of searching through folders or wondering who to ask, they can say things like, “I’m meeting with this customer tomorrow. What should I know?”, “Has this location had staffing problems before?”, “Why do we handle this account differently?”, or “What should I watch for during my first month?” The GPT is now being used by the incoming manager as an ongoing reference tool. The result is essentially a searchable, interactive version of the institutional knowledge that previously would have disappeared with the departing employee. It reduces the “I don’t know what I don’t know” period of starting a new position, shortens the learning curve, improves operational continuity and helps prevent the next person from having to relearn years of lessons through trial and error. The same workflow could be recreated for retiring executives, salespeople, plant managers, administrators, project managers, maintenance supervisors or anyone whose experience contains valuable institutional knowledge. The employee can leave. Their knowledge doesn’t have to.

Instead of relying on a single AI chat, I structure the work as an executive team of specialist perspectives. I define the problem, assign roles such as CTO, product leader, economist, marketer, and skeptic, and have each expert analyze it independently. Then I run several rounds of critique so they can challenge one another’s assumptions before synthesizing their areas of agreement and disagreement. The process ends with a recommendation that includes the rationale and any remaining uncertainties. Step-by-step: 1. Define the problem clearly. 2. Assign specialist roles, such as CTO, product leader, economist, marketer, and skeptic. 3. Have each expert analyze the problem independently. 4. Run several rounds of critique in which the experts challenge one another’s assumptions. 5. Synthesize the areas of agreement and disagreement. 6. Produce a final recommendation with its rationale and remaining uncertainties.

Most days, I want one capable general AI agent. The failure mode appears with higher-stakes questions in security, finance, law, or medicine: my main chat is already full of project context, the agent may answer confidently outside its depth, and a “specialist” prompt in the same thread still inherits the assumptions and tools that session has already loaded. I set up Hermes Agent with a consultation workflow that cold-starts an isolated specialist profile for a one-shot answer. I keep four advisors ready: security, financial, legal, and medical. Each call starts a separate process with its own model, tools, and skills. The specialist has no access to my parent conversation, so I have to write a self-contained prompt. It takes anywhere from a few seconds to a couple of minutes—not instantly—which is fine because I only use it when the stakes justify the wait. The gate is the part I care about most. Nothing runs until I explicitly name the profile, approve the exact prompt, and confirm. There is no silent auto-routing. When I want a second opinion, I draft my own take first, send the same question to the specialist, and compare where we agree and differ. This keeps the general agent honest and prevents specialty advice from becoming the default for routine work. I also keep the scope tight. I limit the medical profile to personal or educational questions, or fully de-identified material—not identifiable patient care. Routine work that my main agent already handles well never goes through consultation. I reserve the specialist profiles for deliberate second opinions or questions I do not want the default session answering alone. Step-by-step: 1. I work through the problem in my main Hermes session and write my own answer first when I want a second opinion. 2. If the question is high-stakes in security, finance, law, or medicine—or I simply want a deliberate second look—I choose the specialist profile to use. 3. I write a self-contained prompt with every fact and constraint the specialist needs because it cannot see the parent chat. 4. I name the profile, approve the prompt, and confirm before anything runs. 5. I let Hermes cold-start that profile as a separate process and return a one-shot answer. 6. I compare my draft with the specialist’s response: where we agree, where we differ, and what still needs a human professional. 7. I skip consultation for routine work and keep medical use limited to personal, educational, or de-identified material.

I built a meal tracker where my only job is to photograph a plate and send it to an agent. Identification, portion estimates, macros, storage, corrections, and the weekly review happen without me. The part worth stealing isn't the food. The agent has its own email address and iMessage thread, so anything I can send from my phone becomes an input. I send a photo through whichever channel takes fewer taps. I send corrections as plain English with `CORRECTION` in the subject, and answer questions in the same thread. There’s no app and no form. Adding a channel took an afternoon and changed what the agent could be pointed at—the tracker is just one use of an agent you can talk to. It polls on a schedule instead of using a webhook because my laptop sleeps and a local gateway would be unavailable half the time. Storage is append-only JSONL. Corrections append a new record with the same ID using last-write-wins, so the stats layer sees one meal while the original estimate stays on disk. Every number can be traced back to the model’s first guess and my override. The accuracy mechanism is a challenge step. Every photo is estimated twice: first by the main model, then by a blind subagent that receives only the photo and the rubric—not the first answer. If it identifies different food, the entry is downgraded to low confidence and becomes a question for me. Testing before making the system autonomous caught bugs that otherwise would have failed silently. The API returns attachments under a different field than I had assumed, and only on the detail endpoint. As a result, a photo email was read as having no attachments, and every meal photo would have been dropped while the daily task logged a tidy “no new meals” and appeared healthy. The send path was separately broken, which would have killed the weekly review on a Sunday even with everything upstream working. A silent no-op that produces a plausible clean run is what this category of build is prone to. The bug that actually cost me was token usage: each blind challenge read used about 44,000 tokens, so a five-photo dinner cost roughly 220,000 tokens for one meal and capped my usage. The fixes, in order of effect, were to run the challenger on a cheap model, skip it once a named product or my confirmation has settled the entry, downscale photos for viewing, and republish the dashboard only when the data changes. Measuring first mattered—I would have blamed photo size, but that was the smaller half. The honest limits are important: calorie estimates from a photo are 20–25% off at best, identification error is a bigger risk than arithmetic error, and alcohol, water, and caffeine are never estimated from photos. The sequence in which I asked for things mattered more than any single instruction: Step-by-step: 1. I gave the goal and the one ingestion mechanic I was sure about, then insisted on an agreed plan before any code. Arguing about storage and failure modes is cheap before code is attached. 2. Before scheduling anything, I processed one real input end to end in front of me, including every outbound path. Inbound gets tested because I use it; the reply and the weekly digest do not. 3. When the agent reported something about my own input that I knew was wrong, I said so and made it re-check. Confidently wrong answers about things I witnessed were the cheapest bugs to find. 4. I asked for an independent second assessment of anything estimated rather than read, and decided up front what level of agreement was enough to accept the result. 5. I asked for a visible audit surface showing every field the agent claims to track, with a correction control on it. 6. I asked the agent to look up anything knowable rather than estimate it. A named product is a lookup; only the unnameable needs a guess. 7. I treated approval friction and token cost as requirements rather than complaints, and measured before changing anything. Almost every rule exists because something went wrong in ordinary use, not because it was designed up front. An agent I can email or text, which keeps records and answers back, is general-purpose; I’ve pointed it at one narrow job. What else would you point it at?

I’m a commercial excellence consultant for industrial B2B businesses, and very few of the clients I speak with have a current view of the competition in their market. Nobody has mapped the competitive landscape recently—sometimes they never have. Clients can often name the companies in the market, but they can’t explain what those competitors do differently or why a customer would choose one over them. That gap is expensive. The mid-market manufacturers I work with, typically in the $75 million to $300 million range, usually don’t have a strategy team to close it. The alternatives are consulting rates for research that goes stale as soon as it’s delivered or a generic template that says the same five things about every market. Neither option is necessarily wrong, but neither is very useful. So I built a Claude Skill: an instruction set that runs live research every time instead of pattern-matching to a generic answer. The input is simple: the company name, its brand voice, the research scope, the geography, and a known competitor list if one is available. Before researching a single competitor, the Skill states the decision the report needs to inform—for example, whether to enter a vertical, how to price against a rival, or where to direct sales next quarter. It checks the client’s own website and capabilities next; an early version once recommended something the client already had. It then maps the competitive field across five tiers: direct, adjacent, disruptor, new entrant, and aspirational for PE-backed clients. It checks Asia-Pacific specifically because that’s the blind spot I’ve seen missed most often. Market size carries a confidence flag instead of false precision. Findings become battlecards phrased the way reps actually talk, rather than in analyst language. The report also includes a threat ranking with a timeline attached, then closes with three opportunities, three risks, and four to six moves for the quarter. Every recommendation is filtered through the original decision instead of being included simply because the research was interesting. The output is two files: an interactive HTML report and a matching PDF. Both are built entirely in HTML and CSS rather than with canvas charts, which can break in exactly the ways that matter—blank on load or missing from the PDF. None of what makes this useful is the AI itself. The important work is naming the decision before researching a competitor, checking what the client already has, tagging confidence instead of faking precision, and writing like a rep rather than an analyst. That’s the difference between a report that gets skimmed once and one that gets acted on. This is one piece of a bigger system I run for industrial manufacturers applying AI to their commercial function. I teach the underlying version of this workflow live. Step-by-step: 1. I provide Claude Skill with the company name, brand voice, research scope, geography, and known competitor list, if available. 2. I define the decision the report needs to inform, such as entering a vertical, pricing against a rival, or directing sales next quarter. 3. I have the Skill check the client’s website and capabilities before researching competitors. 4. I map the competitive field across direct, adjacent, disruptor, new entrant, and, for PE-backed clients, aspirational competitors. 5. I check Asia-Pacific specifically to address a commonly missed blind spot. 6. I assign confidence flags to market-size estimates rather than presenting false precision. 7. I turn the findings into sales-rep-friendly battlecards and add a threat ranking with a timeline. 8. I close the report with three opportunities, three risks, and four to six quarterly moves tied back to the original decision. 9. I deliver the result as an interactive HTML report and a matching PDF, using HTML and CSS instead of canvas charts.

I built an archery app that uses AI to detect arrows and bullseyes on an archery target. It groups the arrows, measures how tight the groupings are, and calculates each arrow’s distance from the bullseye. It also shows the arrows’ locations and their relationship to the bullseye—for example, whether a shot is too far left, right, high, or low, or is dead on. The app can use targets from competition standings to score a shoot according to different standards. After shooting a set of arrows, the archer takes a photo, and the AI detects the target, identifies one or more bullseyes, predicts which arrows are intended for each target, and completes the measurements and scoring almost instantly. The results can then be sent to a coach, who can provide feedback, tips, and techniques to help improve the archer’s shooting. I trained my own model using 3,000 photographs that I took and hand-labeled with the bullseyes and arrows identified. I ran a series of training sessions over several weeks and refined the model to improve its accuracy. It currently achieves about 95% accuracy for arrows and about 90% accuracy for bullseyes. Step-by-step: 1. I took 3,000 photographs of archery targets. 2. I hand-labeled the arrows and bullseyes in those photographs. 3. I trained my own AI model in a series of sessions over several weeks. 4. I refined the model to improve its detection accuracy. 5. An archer shoots a set of arrows and takes a photo of the target. 6. The app detects the target, one or more bullseyes, and the arrows, then predicts which arrows are intended for each target. 7. The app groups the arrows, measures grouping tightness and distance from the bullseye, identifies each arrow’s position relative to the bullseye, and scores the shoot according to the selected standard. 8. The results are sent to a coach for feedback and advice on improving the archer’s shooting.

My doctor asked me to track my blood pressure for a month because it was on the high side before prescribing any medication. I initially recorded each reading manually in an Excel sheet, but after a few days, I wanted a simpler way to enter and manage the data. I uploaded the sheet to Claude Code and asked it to build a blood pressure tracking app. After the app was built, I hosted it on Netlify, used Supabase as the backend to save data for both my wife and me, and added it to my iPhone Home Screen. Step-by-step: 1. I started tracking my blood pressure in an Excel sheet as my doctor requested. 2. After several days of entering the readings manually, I uploaded the sheet to Claude Code. 3. I asked Claude Code to build an app for tracking blood pressure. 4. I hosted the app on Netlify. 5. I used Supabase as the backend to save blood pressure data for both my wife and me. 6. I saved the app to my iPhone Home Screen for easier access.