[{"data":1,"prerenderedAt":552},["ShallowReactive",2],{"blog-list-en":3},[4,129,209,293,373,453],{"id":5,"title":6,"author":7,"body":8,"category":111,"cover":112,"date":113,"description":114,"draft":115,"extension":116,"meta":117,"navigation":118,"path":119,"readingTime":120,"seo":121,"stem":122,"tags":123,"__hash__":128},"blog_en\u002Fen\u002Fblog\u002Fflowbit-mcp-ile-otomatik-gorev-yonetimi.md","FlowBit and MCP: automated task management that works","Codifya",{"type":9,"value":10,"toc":102},"minimark",[11,15,20,23,26,30,33,46,50,62,65,77,80,84,95],[12,13,14],"p",{},"Shipping yet another project management tool meant answering two questions: where to put AI, and where not to. FlowBit is the product of those two decisions.",[16,17,19],"h2",{"id":18},"why-another-pm-tool","Why another PM tool",[12,21,22],{},"Most teams using Jira, Linear, or Asana kept saying the same thing: the tool can do too much, and the team does too little of it. Complexity slows small teams down and forces large teams to rewrite their own processes.",[12,24,25],{},"When designing FlowBit we prioritized: few but right features, and AI assistance exactly where users touch the product daily.",[16,27,29],{"id":28},"why-mcp-matters","Why MCP matters",[12,31,32],{},"MCP (Model Context Protocol) creates a standard bridge between AI models and applications:",[34,35,36,40,43],"ul",{},[37,38,39],"li",{},"Claude, GPT-4, or a local model (Ollama) connect through one integration.",[37,41,42],{},"The model can see in-app context (team, sprint, ticket) without separate system prompts.",[37,44,45],{},"Data scoping is explicit — the model only reads what's allowed.",[16,47,49],{"id":48},"what-we-observed-in-the-field","What we observed in the field",[12,51,52,53,57,58,61],{},"After three months with pilot teams, the clearest finding: AI adds the most value not at the ",[54,55,56],"em",{},"deciding"," moment, but at the ",[54,59,60],{},"remembering"," one.",[12,63,64],{},"A typical flow:",[66,67,68,71,74],"ol",{},[37,69,70],{},"Meeting notes get transcribed.",[37,72,73],{},"FlowBit suggests likely tasks and owners from the text.",[37,75,76],{},"Sprint reports — who did what, where it stalled — drop into Slack automatically.",[12,78,79],{},"None of these steps is \"smart\" individually. Together, they erase the team's \"we forgot about that\" moments.",[16,81,83],{"id":82},"what-we-learned","What we learned",[34,85,86,89,92],{},[37,87,88],{},"Auto-suggested tasks must always pass through human approval. AI shouldn't create tickets alone.",[37,90,91],{},"Reports should be short. Anything past 8 lines doesn't get read.",[37,93,94],{},"Local mode (Ollama) is in demand — especially in the public and healthcare sectors.",[12,96,97,98,101],{},"FlowBit is still early. But positioning AI as a ",[54,99,100],{},"reminder"," rather than an accelerator is working.",{"title":103,"searchDepth":104,"depth":104,"links":105},"",3,[106,108,109,110],{"id":18,"depth":107,"text":19},2,{"id":28,"depth":107,"text":29},{"id":48,"depth":107,"text":49},{"id":82,"depth":107,"text":83},"FlowBit",null,"2026-05-08","The useful part of AI-assisted task creation is making it easier for the team to decide where to focus. FlowBit's design choices and what we observed in the field.",false,"md",{},true,"\u002Fen\u002Fblog\u002Fflowbit-mcp-ile-otomatik-gorev-yonetimi",6,{"title":6,"description":114},"en\u002Fblog\u002Fflowbit-mcp-ile-otomatik-gorev-yonetimi",[124,125,126,127],"ai","mcp","product management","automation","eEPaJPHkXnNQIk4MQTX7y9-HiSS6VIOKRQGV8lN51P0",{"id":130,"title":131,"author":7,"body":132,"category":195,"cover":112,"date":196,"description":197,"draft":115,"extension":116,"meta":198,"navigation":118,"path":199,"readingTime":200,"seo":201,"stem":202,"tags":203,"__hash__":208},"blog_en\u002Fen\u002Fblog\u002Fbarbershop-appointment-intervals.md","What we learned from 850 barbershops about appointment intervals",{"type":9,"value":133,"toc":190},[134,141,145,152,159,163,166,180,183,187],[12,135,136,137],{},"We examined appointment data from over 850 barbershops using BerberBul. The question was simple: ",[138,139,140],"strong",{},"what is the most productive appointment interval?",[16,142,144],{"id":143},"findings","Findings",[12,146,147,148,151],{},"The highest average occupancy rate was among barbers working with ",[138,149,150],{},"35-minute intervals",". Shops using intervals under 30 minutes saw lower customer satisfaction — especially with services that include hair washing. Shops with intervals above 45 minutes had reduced daily capacity.",[12,153,154,155,158],{},"Unexpected finding: ",[138,156,157],{},"the day before Sunday"," (Saturday) was the busiest. But the most productive day was Tuesday. Tuesday's average occupancy was 82%, while Saturday was 67% — busier but less profitable because wait times stretched.",[16,160,162],{"id":161},"what-we-recommend","What we recommend",[12,164,165],{},"Two solutions for Saturday:",[66,167,168,174],{},[37,169,170,173],{},[138,171,172],{},"Move earlier appointments to Saturday morning"," — 09:00-11:00 has the highest booking rate.",[37,175,176,179],{},[138,177,178],{},"Use time blocks"," — instead of 30-minute slots, use 45-minute \"peak period blocks.\"",[12,181,182],{},"Barber shops using BerberBul's automatic conflict detection with this kind of planning see 23% lower cancellation rates.",[16,184,186],{"id":185},"the-real-lesson-from-the-data","The real lesson from the data",[12,188,189],{},"Appointment system design varies by barbershop physical capacity and staff experience. There's no single right answer; but with the right data, each business can find its own. BerberBul makes this measurable.",{"title":103,"searchDepth":104,"depth":104,"links":191},[192,193,194],{"id":143,"depth":107,"text":144},{"id":161,"depth":107,"text":162},{"id":185,"depth":107,"text":186},"BerberBul","2026-05-02","We analyzed appointment behavior from 850+ barbershops on BerberBul. The optimal interval turned out to be different from what most shops assumed.",{},"\u002Fen\u002Fblog\u002Fbarbershop-appointment-intervals",5,{"title":131,"description":197},"en\u002Fblog\u002Fbarbershop-appointment-intervals",[204,205,206,207],"data","operations","scheduling","insights","ag4fwhZ8V1QMliIrvt7N4fp8E0Bt3ohDCPlqN42xjNY",{"id":210,"title":211,"author":7,"body":212,"category":280,"cover":112,"date":281,"description":282,"draft":115,"extension":116,"meta":283,"navigation":118,"path":284,"readingTime":285,"seo":286,"stem":287,"tags":288,"__hash__":292},"blog_en\u002Fen\u002Fblog\u002Fcityos-vatandas-bildirim-sistemi.md","Designing a citizen reporting system from scratch",{"type":9,"value":213,"toc":273},[214,217,220,224,227,231,234,241,245,248,255,259,262,266],[12,215,216],{},"Citizen reporting is one of the most visible service metrics for a municipality — and usually one of the weakest operationally. Call center, social media, mobile app: each channel tends to live in its own silo.",[12,218,219],{},"When designing CityOS's reporting module, we set out to solve three core problems.",[16,221,223],{"id":222},"_1-a-unified-queue","1. A unified queue",[12,225,226],{},"Which channel a citizen used — phone, mobile app, web — should be a detail the assignee doesn't have to think about. CityOS pipes every channel into one queue; the assigned report looks the same regardless of source.",[16,228,230],{"id":229},"_2-auto-classification-but-overridable","2. Auto-classification, but overridable",[12,232,233],{},"Whether a report is \"park maintenance\" or \"road maintenance\" is obvious in roughly 80% of cases. The remaining 20% needs a human.",[12,235,236,237,240],{},"CityOS auto-classifies the incoming report, but the assignee can always change the category. That small detail — ",[54,238,239],{},"\"AI suggested it, but you have the final say\""," — dramatically increases acceptance.",[16,242,244],{"id":243},"_3-live-updates-for-the-citizen","3. Live updates for the citizen",[12,246,247],{},"When the internal process completes, push notifications are the fastest way to tell the citizen. But which step should we notify? Assigned, in-progress, completed?",[12,249,250,251,254],{},"Three pilot municipalities tested it and the result was clear: a single notification — ",[54,252,253],{},"\"Your report is complete\""," — produced higher satisfaction than multiple updates. Notifying every step felt like nagging.",[16,256,258],{"id":257},"the-invisible-part-gdprkvkk","The invisible part: GDPR\u002FKVKK",[12,260,261],{},"Separating identity from report content, enabling audit log from day one, and minimizing personal data — all standard. This is the most expensive layer to add later.",[16,263,265],{"id":264},"takeaway","Takeaway",[12,267,268,269,272],{},"A reporting system requires far more decisions than it appears to. CityOS's module is a concrete application of the ",[54,270,271],{},"less clicking, more clarity"," principle.",{"title":103,"searchDepth":104,"depth":104,"links":274},[275,276,277,278,279],{"id":222,"depth":107,"text":223},{"id":229,"depth":107,"text":230},{"id":243,"depth":107,"text":244},{"id":257,"depth":107,"text":258},{"id":264,"depth":107,"text":265},"CityOS","2026-04-22","What questions did we ask while designing CityOS's reporting module? Three points where municipal workflows tend to break, and the answers we built.",{},"\u002Fen\u002Fblog\u002Fcityos-vatandas-bildirim-sistemi",7,{"title":211,"description":282},"en\u002Fblog\u002Fcityos-vatandas-bildirim-sistemi",[289,290,291],"public sector","product design","workflow","NPgZezqxZi8KgQ0XrRLvamU8IRsYrBEWgcwiJc-PNbg",{"id":294,"title":295,"author":7,"body":296,"category":280,"cover":112,"date":362,"description":363,"draft":115,"extension":116,"meta":364,"navigation":118,"path":365,"readingTime":285,"seo":366,"stem":367,"tags":368,"__hash__":372},"blog_en\u002Fen\u002Fblog\u002Fcityos-40000-reports-case.md","CityOS in a district of 180,000: 6 months, 40,000 tickets",{"type":9,"value":297,"toc":356},[298,301,305,308,311,315,318,321,325,328,334,340,346,350,353],[12,299,300],{},"In January 2026, a district in Antalya deployed CityOS. By June, over 40,000 reports had been received and resolved. Beyond the numbers, the most valuable lessons from this period:",[16,302,304],{"id":303},"process-first-software-second","Process first, software second",[12,306,307],{},"The mayor's first question in the initial meeting wasn't about software features: \"How do we not reduce the permanent staff count?\" This was the right question. If they'd adopted CityOS just to \"collect reports faster,\" it would have been the same as the old system.",[12,309,310],{},"The solution: instead of an agent chasing reports, build a team that meets reports in the field. 12 street representatives in the city center, each responsible for 4 neighborhoods. Every morning their mobile app shows daily routes.",[16,312,314],{"id":313},"finding-category-matters-more-than-volume","Finding: category matters more than volume",[12,316,317],{},"In the first two months, the top report type was \"pothole.\" An analysis revealed that 600 of 1,000 potholes were in the same heatmap — an area with heavy winter rainfall and old infrastructure.",[12,319,320],{},"When CityOS's reporting view started showing the mayor \"the source of the problem,\" it transformed from a routine request handler into a decision-support tool.",[16,322,324],{"id":323},"three-slowdown-points","Three slowdown points",[12,326,327],{},"As with any system, three points caused delays:",[12,329,330,333],{},[138,331,332],{},"1. Vehicle assignment:"," Directing a report to the field is the most critical step. Average wait time was 4 hours. A rotation algorithm was developed to bring this down to 90 minutes.",[12,335,336,339],{},[138,337,338],{},"2. Citizen feedback:"," There was no \"what happened\" response to the citizen who filed the report. A closed-loop notification was added: \"Your request was received → team dispatched → resolved.\"",[12,341,342,345],{},[138,343,344],{},"3. Neighborhood headmen (muhtars):"," The headmen were outside the system. After giving them read access, they started checking before complaints even arrived.",[16,347,349],{"id":348},"outcome","Outcome",[12,351,352],{},"Of 40,000 reports in 6 months, 94% were resolved within 14 days. The municipality's own data shows this was 61% in the old system. The difference wasn't in the technology; it was in how the team receiving reports was designed.",[12,354,355],{},"CityOS became a tool that helped the municipality rethink its own internal processes. The software didn't solve the problem — it made the problem visible.",{"title":103,"searchDepth":104,"depth":104,"links":357},[358,359,360,361],{"id":303,"depth":107,"text":304},{"id":313,"depth":107,"text":314},{"id":323,"depth":107,"text":324},{"id":348,"depth":107,"text":349},"2026-04-20","How a district municipality managed 40,000 citizen reports in 6 months with CityOS. The design decisions that determined success — not the technology.",{},"\u002Fen\u002Fblog\u002Fcityos-40000-reports-case",{"title":295,"description":363},"en\u002Fblog\u002Fcityos-40000-reports-case",[369,370,371,205],"case study","municipality","citizen engagement","Nd-D3hXbvFQLP_puK3e6rgBun6y2UllB94Ph5wgmAaI",{"id":374,"title":375,"author":7,"body":376,"category":443,"cover":112,"date":444,"description":445,"draft":115,"extension":116,"meta":446,"navigation":118,"path":447,"readingTime":200,"seo":448,"stem":449,"tags":450,"__hash__":452},"blog_en\u002Fen\u002Fblog\u002Fqr-menu-kafe-vakasi.md","A café's switch to QR menu: 90 days later",{"type":9,"value":377,"toc":437},[378,381,401,404,408,411,415,418,422,429,432,434],[12,379,380],{},"In March, a single-location café in Konya switched to QR Menu. Ninety days in, here's what they shared:",[34,382,383,389,395],{},[37,384,385,388],{},[138,386,387],{},"90% drop"," in printed-menu cost (only the table cards get reprinted)",[37,390,391,394],{},[138,392,393],{},"60% fewer"," order errors from international customers (thanks to auto-translation)",[37,396,397,400],{},[138,398,399],{},"8 minutes saved per day"," on menu updates",[12,402,403],{},"More interesting than the numbers were the behavior shifts.",[16,405,407],{"id":406},"how-long-do-customers-stay-on-the-menu","How long do customers stay on the menu?",[12,409,410],{},"Average viewing time went from one minute to 2.5. We initially read this as a bad signal — longer might mean more indecision. According to surveys, the opposite was true: customers were now reviewing photos, checking allergen labels, and reading content details.",[16,412,414],{"id":413},"the-waiters-role-changed","The waiter's role changed",[12,416,417],{},"The pressure on the waiter to answer \"what do you recommend?\" eased, because the customer had already seen detailed info on the menu. That freed the waiter to focus on table atmosphere and service.",[16,419,421],{"id":420},"the-unexpectedly-popular-feature","The unexpectedly popular feature",[12,423,424,425,428],{},"The owner did not predict that ",[138,426,427],{},"stock-based auto-hiding"," would matter most. Sold-out items quietly disappear from the menu — the customer orders, the waiter never has to apologize.",[12,430,431],{},"That single detail meaningfully improved customer satisfaction.",[16,433,265],{"id":264},[12,435,436],{},"Switching to QR menu isn't just a paper-to-digital migration. Done right, it's an infrastructure that reduces operational load and errors while quietly improving the customer experience.",{"title":103,"searchDepth":104,"depth":104,"links":438},[439,440,441,442],{"id":406,"depth":107,"text":407},{"id":413,"depth":107,"text":414},{"id":420,"depth":107,"text":421},{"id":264,"depth":107,"text":265},"QR Menü","2026-03-30","What changed for a single-location café after switching to QR Menu? A short case study on operations, customer behavior, and error rate.",{},"\u002Fen\u002Fblog\u002Fqr-menu-kafe-vakasi",{"title":375,"description":445},"en\u002Fblog\u002Fqr-menu-kafe-vakasi",[369,451,205],"restaurant","6ZxiVmPuaiJF0I3FvOYwjKItZUX8Lo9I6id0kwBEQlY",{"id":454,"title":455,"author":7,"body":456,"category":540,"cover":112,"date":541,"description":542,"draft":115,"extension":116,"meta":543,"navigation":118,"path":544,"readingTime":120,"seo":545,"stem":546,"tags":547,"__hash__":551},"blog_en\u002Fen\u002Fblog\u002Fcodifya-vizyon-sade-mimari.md","Calm architecture: fewer concepts, fewer clicks, more outcomes",{"type":9,"value":457,"toc":534},[458,464,470,474,492,495,499,506,509,513,524,528,531],[12,459,460,461],{},"Over the years we noticed: the reason a user abandons an app is rarely a missing feature. More often it's the exhaustion of an interface that ",[54,462,463],{},"does many things but explains none clearly.",[12,465,466,467],{},"At Codifya every product follows one principle: ",[138,468,469],{},"fewer concepts, fewer clicks, more outcomes.",[16,471,473],{"id":472},"fewer-concepts","\"Fewer concepts\"",[12,475,476,477,480,481,484,485,480,488,491],{},"The more new terms a user has to learn, the longer they feel like a beginner. In CityOS we use ",[54,478,479],{},"\"report\""," instead of ",[54,482,483],{},"\"ticket\""," and ",[54,486,487],{},"\"owner\"",[54,489,490],{},"\"agent\""," — because those words already exist in municipal vocabulary.",[12,493,494],{},"In FlowBit, terms like \"epic\" and \"story point\" are hidden by default. Teams that need them turn them on; most teams just look at \"task\" and \"sprint.\"",[16,496,498],{"id":497},"fewer-clicks","\"Fewer clicks\"",[12,500,501,502,505],{},"If a user takes four clicks to reach a goal, we look for the path to three. This is not about speed — it's about ",[54,503,504],{},"focus",". Every click is a small distracting decision.",[12,507,508],{},"In QR Menu, table number is auto-detected (embedded in the QR), language is auto-detected (from browser locale), and order confirmation is one tap. None of this is visible — but it's felt.",[16,510,512],{"id":511},"more-outcomes","\"More outcomes\"",[12,514,515,516,519,520,523],{},"Calm doesn't mean ",[54,517,518],{},"doing less","; it means ",[54,521,522],{},"appearing to do less",". The difference is in the infrastructure. BerberBul's booking system looks calm; behind it run conflict checks, auto-cancellations, and push reminders.",[16,525,527],{"id":526},"conclusion","Conclusion",[12,529,530],{},"A calm product comes from a calm architecture. A calm architecture is the byproduct of engineering culture: choosing the right abstraction, refusing unnecessary layers, and questioning every line of code.",[12,532,533],{},"Codifya's four products are the concrete expression of that culture.",{"title":103,"searchDepth":104,"depth":104,"links":535},[536,537,538,539],{"id":472,"depth":107,"text":473},{"id":497,"depth":107,"text":498},{"id":511,"depth":107,"text":512},{"id":526,"depth":107,"text":527},"Vizyon","2026-03-12","The principles behind every Codifya product. Calm is not an aesthetic preference — it's the precondition for sustainability.",{},"\u002Fen\u002Fblog\u002Fcodifya-vizyon-sade-mimari",{"title":455,"description":542},"en\u002Fblog\u002Fcodifya-vizyon-sade-mimari",[548,549,550],"design philosophy","simplicity","engineering culture","gOyTWPil0kftqVFgqXx5Hnf2_fgW3DWpGx61vqXOpkE",1781523770662]