[{"data":1,"prerenderedAt":1254},["ShallowReactive",2],{"blog-articles-for-slug-conversational-ai-design-why-bad-bots-are-a-design-problem":3},{"articles":4,"publishedKeys":1253},[5,25,40,58,72,89,104,120,133,152,169,184,202,218,236,253,272,286,302,321,335,348,710,724,736,751,764,780,795,807,820,836,853,871,887,902,919,934,947,959,973,986,1002,1016,1030,1043,1056,1071,1086,1101,1116,1130,1145,1159,1176,1194,1211,1225,1240],{"id":6,"title":7,"author":8,"category":9,"date":10,"description":11,"extension":12,"featured":13,"image":14,"meta":15,"metaDescription":16,"metaTitle":17,"navigation":18,"path":19,"seo":20,"stem":21,"tags":22,"__hash__":23,"__readTime":24},"blog\u002Fblog\u002F15-advanced-features-that-turn-your-chatbots-into-gold.md","15 advanced features that turn your chatbots into gold","Len Debets","Product Features","2020-08-20","A chatbot that answers a few FAQs is easy to build. Here are 15 advanced features that move your chatbot from a simple Q&A tool to something your business depends on.","md",false,"\u002Fblog\u002F3327590-300x200.jpg",{},"A chatbot that answers a few FAQs is easy to build. These 15 advanced features turn it into something your business actually depends on.","15 Advanced Chatbot Features Worth Building",true,"\u002Fblog\u002F15-advanced-features-that-turn-your-chatbots-into-gold",{"title":7,"description":11},"blog\u002F15-advanced-features-that-turn-your-chatbots-into-gold",[],"2Ub1RORB7pfLmgONRupxdg1IOzCXXrqje5046hBrjbA","3 min read",{"id":26,"title":27,"author":8,"category":28,"date":29,"description":30,"extension":12,"featured":13,"image":31,"meta":32,"metaDescription":33,"metaTitle":34,"navigation":18,"path":35,"seo":36,"stem":37,"tags":34,"__hash__":38,"__readTime":39},"blog\u002Fblog\u002F9-things-i-hate-about-ai.md","9 Things I Really Hate About AI","AI Technology","2025-05-12","I think it's great that technology is so embedded in our daily lives. It helps us get knowledge faster, finish tasks more efficiently, gives us inspiration, and occasionally scares the hell out of us with those crazy (fake) videos. I help a lot of companies implement AI, so in the end it pays my bills. But after spending a ridiculous amount of time with all these new technologies, it's time to write down the things I really hate about AI.","\u002Fblog\u002F1747039600744.png",{},"I help companies implement AI and it pays my bills. After a ridiculous amount of time with these tools, here are the 9 things I really hate about AI.",null,"\u002Fblog\u002F9-things-i-hate-about-ai",{"title":27,"description":30},"blog\u002F9-things-i-hate-about-ai","x3Pfa3upWCZeJg3PuHjgqCGhd8w1PdeVtVLlYXYmDZk","7 min read",{"id":41,"title":42,"author":8,"category":28,"date":43,"description":44,"extension":12,"featured":13,"image":45,"meta":46,"metaDescription":47,"metaTitle":34,"navigation":18,"path":48,"seo":49,"stem":50,"tags":51,"__hash__":57,"__readTime":39},"blog\u002Fblog\u002Fa-conversation-is-not-a-case.md","A Conversation Is Not a Case","2026-10-06","The 95 percent pilot failure stat is making the rounds again, this time with a diagnosis I actually agree with: chatbots automate sessions, but a bank runs cases. Why the unit of work decides whether your AI survives production, and the day eight test that shows which one you built.","\u002Fblog\u002Fblits-session-vs-case-hero.png",{},"Chatbots automate sessions, but a bank runs cases. Why the unit of work decides whether your AI survives production, and the day eight test.","\u002Fblog\u002Fa-conversation-is-not-a-case",{"title":42,"description":44},"blog\u002Fa-conversation-is-not-a-case",[52,53,54,55,56],"agentic ai","enterprise ai","ai strategy","case management","banking","IJO6dpCGt7jjLgV0FvRqM03LiuYvz779pb3P824edrk",{"id":59,"title":60,"author":8,"category":61,"date":62,"description":63,"extension":12,"featured":13,"image":64,"meta":65,"metaDescription":66,"metaTitle":34,"navigation":18,"path":67,"seo":68,"stem":69,"tags":34,"__hash__":70,"__readTime":71},"blog\u002Fblog\u002Fa-day-in-the-life-of-ai-scale-up.md","A day in the life of an AI scale-up","Company News","2025-02-24","AI moves at breakneck speed, and being the CTO of a fast growing scaleup in it is a great job. It is also not for the faint of heart. We work on large language models (LLMs), chatbots, digital humans and voicebots, and most of our customers are large banks, international companies and professional enterprises.","\u002Fblog\u002F1740065406903.png",{},"What it is like to be CTO of a fast growing AI scaleup that builds LLMs, chatbots, digital humans and voicebots for banks and large enterprises.","\u002Fblog\u002Fa-day-in-the-life-of-ai-scale-up",{"title":60,"description":63},"blog\u002Fa-day-in-the-life-of-ai-scale-up","TiXNFQdXhajzvfAWWOBvKGpeG5WGq1A5v2MXtyvF-w0","5 min read",{"id":73,"title":74,"author":8,"category":28,"date":75,"description":76,"extension":12,"featured":13,"image":77,"meta":78,"metaDescription":34,"metaTitle":79,"navigation":18,"path":80,"seo":81,"stem":82,"tags":83,"__hash__":87,"__readTime":88},"blog\u002Fblog\u002Fagentic-ai-dialects-and-the-voice-quality-gap.md","Agentic AI Languages and Dialects: Why Voice Quality Is Still the Hard Part","2026-04-10","Agentic systems promise autonomous workflows, but speech and dialect quality fail first. Why Arabic, Turkish, Chinese and more run into the same gap.","\u002Fblog\u002Fdialect.jpg",{},"Agentic AI Dialects and the Voice Quality Gap","\u002Fblog\u002Fagentic-ai-dialects-and-the-voice-quality-gap",{"title":74,"description":76},"blog\u002Fagentic-ai-dialects-and-the-voice-quality-gap",[52,84,85,86,53],"speech ai","dialects","multilingual ai","e0ZVAmJonwtrh9SFlhMirDWXgPrHQL-HItCKsC7cOFA","14 min read",{"id":90,"title":91,"author":8,"category":9,"date":92,"description":93,"extension":12,"featured":13,"image":94,"meta":95,"metaDescription":34,"metaTitle":34,"navigation":18,"path":96,"seo":97,"stem":100,"tags":101,"__hash__":102,"__readTime":103},"blog\u002Fblog\u002Fagentic-ai-studio-for-enterprises.md","Introducing the Agentic AI Studio for Enterprises","2026-02-17","Introducing Agentic AI Studio: a platform agnostic, enterprise ready way to design and run autonomous workflows with your own models and data.","\u002Fimages\u002Fblits-agentic-studio.png",{},"\u002Fblog\u002Fagentic-ai-studio-for-enterprises",{"title":98,"description":99},"Introducing the Agentic AI Studio for Enterprise","Self running AI agents are changing how enterprise processes work. We're releasing the Agentic AI Studio for Blits.ai, a platform agnostic, enterprise ready way to design and run autonomous workflows with your own models and data.","blog\u002Fagentic-ai-studio-for-enterprises",[],"Iko_ojcsirkF69bOvZZlrkHTDh-kRPk1Jr-wTgc8-no","9 min read",{"id":105,"title":106,"author":8,"category":28,"date":107,"description":108,"extension":12,"featured":13,"image":109,"meta":110,"metaDescription":111,"metaTitle":112,"navigation":18,"path":113,"seo":114,"stem":115,"tags":116,"__hash__":119,"__readTime":71},"blog\u002Fblog\u002Fagentic-ai-vs-generative-ai.md","Agentic AI vs Generative AI: What Changes When AI Starts Doing the Work","2026-10-23","Generative AI writes. Agentic AI does. That sounds like a small step, but it changes what can go wrong, what you need to govern and how you measure success. The real difference between agentic AI and generative AI, why most agentic projects stall, and what the ones in production have in common.","\u002Fblog\u002Fblits-agentic-vs-generative-hero.png",{},"Agentic AI vs generative AI explained for enterprises: the real difference, why most agentic projects stall, what BNY runs in production and what to govern.","Agentic AI vs Generative AI: What Really Changes","\u002Fblog\u002Fagentic-ai-vs-generative-ai",{"title":106,"description":108},"blog\u002Fagentic-ai-vs-generative-ai",[52,117,53,54,118],"generative ai","ai governance","0m-VgMpeulu0O6n741Xc_JzXzqKjxZfQNWYe0R9Qjow",{"id":121,"title":122,"author":8,"category":28,"date":123,"description":124,"extension":12,"featured":13,"image":125,"meta":126,"metaDescription":34,"metaTitle":127,"navigation":18,"path":128,"seo":129,"stem":130,"tags":131,"__hash__":132,"__readTime":71},"blog\u002Fblog\u002Fagentic-pay-and-the-moment-ai-was-allowed-to-spend-money-by-itself.md","Agentic Pay and the Moment AI Was Allowed to Spend Money","2026-01-11","Why giving AI agents controlled access to payments changes a lot, and why the open question is who designs the rules they spend under.","\u002Fimages\u002Fblits-agentic-pay.jpg",{},"Agentic Pay: When AI Was Allowed to Spend Money","\u002Fblog\u002Fagentic-pay-and-the-moment-ai-was-allowed-to-spend-money-by-itself",{"title":122,"description":124},"blog\u002Fagentic-pay-and-the-moment-ai-was-allowed-to-spend-money-by-itself",[],"njF5BBkPJMdXpGxOH8XGMC0IglTZeDkrA99y3UB_x3c",{"id":134,"title":135,"author":136,"category":28,"date":137,"description":138,"extension":12,"featured":13,"image":139,"meta":140,"metaDescription":141,"metaTitle":142,"navigation":18,"path":143,"seo":144,"stem":145,"tags":146,"__hash__":150,"__readTime":151},"blog\u002Fblog\u002Fagentic-workflows-explained.md","Agentic Workflows Explained, With Real Examples From Banking and Insurance","Coen Doolaard","2026-10-30","An agentic workflow lets AI work through a task in steps: it plans, uses your systems, checks its own work and hands the important decision to a person. A practical explanation for banks and insurers, with the design patterns, real production examples and a checklist before you build.","\u002Fblog\u002Fblits-agentic-workflows-hero.png",{},"What agentic workflows are, how they differ from AI agents, the four design patterns behind them, and what banks and insurers already run in production.","Agentic Workflows Explained: Examples From Banking and Insurance","\u002Fblog\u002Fagentic-workflows-explained",{"title":135,"description":138},"blog\u002Fagentic-workflows-explained",[147,52,56,148,149],"agentic workflows","insurance","ai architecture","BLwbkWwurQLOnl6HiQG1ydDkDBxpul46cip7v8MTUrE","6 min read",{"id":153,"title":154,"author":8,"category":9,"date":155,"description":156,"extension":12,"featured":13,"image":157,"meta":158,"metaDescription":159,"metaTitle":160,"navigation":18,"path":161,"seo":162,"stem":163,"tags":164,"__hash__":168,"__readTime":103},"blog\u002Fblog\u002Fagents-that-ask-remember-and-show-their-work.md","Agents That Ask, Remember and Show Their Work: What's New in Blits.ai","2026-10-28","Release 5.9 is about one thing: giving AI agents real work without losing control of it. Agents now ask for approval before risky actions, record every decision, publish only after their tests pass, remember returning customers within strict limits and work with other companies' agents over the open A2A protocol. What each feature does, and how it works under the hood.","\u002Fblog\u002Fblits-v59-features-hero.png",{},"Blits.ai release 5.9: approval per tool, a decision log per conversation, tested publishing, memory per customer, the A2A protocol and provider control.","What's New in Blits.ai: Agents That Ask and Show Their Work","\u002Fblog\u002Fagents-that-ask-remember-and-show-their-work",{"title":154,"description":156},"blog\u002Fagents-that-ask-remember-and-show-their-work",[165,52,118,166,167],"product update","human in the loop","a2a","aggHmx6DtKNch6bPEhQ64Sj4iwGK2fcVM5nmTBIO5jc",{"id":170,"title":171,"author":8,"category":28,"date":172,"description":173,"extension":12,"featured":13,"image":174,"meta":175,"metaDescription":176,"metaTitle":171,"navigation":18,"path":177,"seo":178,"stem":179,"tags":180,"__hash__":183,"__readTime":71},"blog\u002Fblog\u002Fai-agency-or-ai-platform.md","AI Agency or AI Platform? How Enterprises Should Choose","2026-11-10","Hire an AI agency, buy a SaaS tool, build it yourself, or work with a platform? Every enterprise faces that choice, and the data is starting to show which path reaches production. Why custom built AI so often gets abandoned, and the questions that tell you whether you're buying a project or a capability.","\u002Fblog\u002Fblits-ai-agency-vs-platform-hero.png",{},"Should you hire an AI agency, buy software or build in house? What the data says about build vs buy, why custom AI gets abandoned, and the questions to ask.","\u002Fblog\u002Fai-agency-or-ai-platform",{"title":171,"description":173},"blog\u002Fai-agency-or-ai-platform",[181,53,182,54,52],"ai agency","build vs buy","3kzZhIokvhURrrQMls8Wry1YOEDElpvbqqdDBOP_BpM",{"id":185,"title":186,"author":136,"category":28,"date":187,"description":188,"extension":12,"featured":13,"image":189,"meta":190,"metaDescription":191,"metaTitle":186,"navigation":18,"path":192,"seo":193,"stem":194,"tags":195,"__hash__":200,"__readTime":201},"blog\u002Fblog\u002Fai-call-center-voice-agents-answer-first.md","AI Call Center: What Changes When Voice Agents Answer First","2026-11-17","An AI call center is not a call center without people. It is one where a voice agent answers first, AI supports the people who handle the harder calls, and every conversation feeds quality. What changes on the floor, what the public record shows, and where the first projects went wrong.","\u002Fblog\u002Fblits-ai-call-center-hero.png",{},"What an AI call center looks like in 2026: voice agents on the first line, AI assist for human agents, every call reviewed, and the mistakes to avoid.","\u002Fblog\u002Fai-call-center-voice-agents-answer-first",{"title":186,"description":188},"blog\u002Fai-call-center-voice-agents-answer-first",[196,197,198,199,56],"ai call center","voice agents","contact center","customer experience","O_O6PWPvMGjX-Ftz9owvzBTO61x4KtOuiRNEk4-Sw_U","8 min read",{"id":203,"title":204,"author":136,"category":28,"date":205,"description":206,"extension":12,"featured":13,"image":207,"meta":208,"metaDescription":209,"metaTitle":210,"navigation":18,"path":211,"seo":212,"stem":213,"tags":214,"__hash__":217,"__readTime":151},"blog\u002Fblog\u002Fai-customer-service-what-actually-works.md","AI Customer Service in 2026: What Actually Works, and What the Benchmarks Hide","2026-10-27","AI customer service works, but not in the way most dashboards suggest. Public deployments show a median containment rate of two thirds, yet contacts drop by only a fifth. What separates the projects that deliver from the ones that quietly get reversed, and what to ask before you buy.","\u002Fblog\u002Fblits-ai-customer-service-hero.png",{},"What AI customer service really delivers in 2026: public benchmarks, the reversals at Klarna and CBA, and what to ask before you buy AI customer service software.","AI Customer Service in 2026: What Actually Works","\u002Fblog\u002Fai-customer-service-what-actually-works",{"title":204,"description":206},"blog\u002Fai-customer-service-what-actually-works",[215,199,198,216,56],"ai customer service","conversational ai","MTic7KjEg4zMr-2JCJeEX9masmR3GM0ZsGw8b4NhwsE",{"id":219,"title":220,"author":136,"category":28,"date":221,"description":222,"extension":12,"featured":13,"image":223,"meta":224,"metaDescription":225,"metaTitle":226,"navigation":18,"path":227,"seo":228,"stem":229,"tags":230,"__hash__":235,"__readTime":71},"blog\u002Fblog\u002Fai-in-insurance-use-cases-in-production.md","AI in Insurance: The Use Cases Already Running in Production","2026-11-03","AI in insurance is no longer an experiment. Of 117 public insurance deployments we track, 103 run in production or at scale. Where insurers use AI today, what the leaders report, what regulators expect, and why most of the industry is still stuck at the pilot stage.","\u002Fblog\u002Fblits-ai-insurance-hero.png",{},"Where insurers already run AI in production: claims, fraud, underwriting and customer service, with results from Allianz, Zurich, Ping An, Lemonade and Aviva.","AI in Insurance: Use Cases Already in Production","\u002Fblog\u002Fai-in-insurance-use-cases-in-production",{"title":220,"description":222},"blog\u002Fai-in-insurance-use-cases-in-production",[231,232,233,52,234],"ai insurance","claims automation","underwriting","customer service","7l-cJ6cSc8Ox8EmQ6T0Q2DdDPy6FW5cLHrD5gx4zvq0",{"id":237,"title":238,"author":8,"category":28,"date":239,"description":240,"extension":12,"featured":13,"image":241,"meta":242,"metaDescription":34,"metaTitle":243,"navigation":18,"path":244,"seo":245,"stem":246,"tags":247,"__hash__":251,"__readTime":252},"blog\u002Fblog\u002Fai-spending-delegation-policies-without-losing-control.md","You Already Know How to Delegate Spending. Now Write the Mandate for an Agent.","2026-03-31","You already delegate spending with cards, POs and dual sign off. An AI agent is a new delegate. How to write its mandate, and where the analogy breaks.","\u002Fimages\u002Fblog-ai-policies.png",{},"Writing the Spending Mandate for an AI Agent","\u002Fblog\u002Fai-spending-delegation-policies-without-losing-control",{"title":238,"description":240},"blog\u002Fai-spending-delegation-policies-without-losing-control",[248,118,53,249,250],"agentic payments","delegation","payments","3UinRxWgwEV2xzGlI2-7RNVqr-WksZKCQs46SjskL6E","11 min read",{"id":254,"title":255,"author":256,"category":28,"date":257,"description":258,"extension":12,"featured":13,"image":259,"meta":260,"metaDescription":261,"metaTitle":262,"navigation":18,"path":263,"seo":264,"stem":265,"tags":266,"__hash__":271,"__readTime":151},"blog\u002Fblog\u002Fai-voice-agents-for-enterprises.md","AI Voice Agents for Enterprises: Latency, Dialects and Compliance","Yuri Ihnatov","2026-11-06","An AI voice agent can answer every call at once, in many languages, at any hour. Whether customers accept it comes down to three things that rarely show up in a demo: how fast it answers, whether it understands the way they actually speak, and whether it follows the rules for disclosure, consent and authentication.","\u002Fblog\u002Fblits-ai-voice-agents-hero.png",{},"What enterprises need from an AI voice agent: response times people accept, speech recognition that works for every dialect, and the rules on disclosure and fraud.","AI Voice Agents for Enterprises: Latency, Dialects, Compliance","\u002Fblog\u002Fai-voice-agents-for-enterprises",{"title":255,"description":258},"blog\u002Fai-voice-agents-for-enterprises",[267,268,198,269,270],"ai voice agents","voice ai","speech recognition","compliance","d_UKYWERVh6gaVcc2b1uSw1fBPQO39q6odNaVlq-jC8",{"id":273,"title":274,"author":275,"category":276,"date":277,"description":278,"extension":12,"featured":13,"image":279,"meta":280,"metaDescription":34,"metaTitle":34,"navigation":18,"path":281,"seo":282,"stem":283,"tags":34,"__hash__":284,"__readTime":285},"blog\u002Fblog\u002Fbenchmarking-conversational-ai-whitepaper.md","Benchmarking Conversational AI: Whitepaper","Paul Coerkamp","Research","2022-08-25","Our research into conversational AI performance shows clear gains in chatbot effectiveness across several industries and use cases.","\u002Fblog\u002FCover-Benchmarking-Conversational-AI-Whitepaper.jpg",{},"\u002Fblog\u002Fbenchmarking-conversational-ai-whitepaper",{"title":274,"description":278},"blog\u002Fbenchmarking-conversational-ai-whitepaper","nfDZr9Jn-Owi4BXRnkyz9owwZBZPJNEZp01hok5_J-Y","2 min read",{"id":287,"title":288,"author":275,"category":28,"date":289,"description":290,"extension":12,"featured":13,"image":291,"meta":292,"metaDescription":34,"metaTitle":293,"navigation":18,"path":294,"seo":295,"stem":296,"tags":297,"__hash__":301,"__readTime":252},"blog\u002Fblog\u002Fbenchmarking-tts-for-customers-beyond-mos-and-demos.md","How to Score a TTS Bake-Off When the Demo Is Lying to You","2026-04-08","MOS has saturated and demos are hand tuned. The lab protocol we use to grade TTS on latency, jitter, entity accuracy, barge in and cloning risk.","\u002Fimages\u002Fblog-benchmark.png",{},"Benchmarking TTS Beyond MOS and Demos","\u002Fblog\u002Fbenchmarking-tts-for-customers-beyond-mos-and-demos",{"title":288,"description":290},"blog\u002Fbenchmarking-tts-for-customers-beyond-mos-and-demos",[298,268,299,300,197],"tts","benchmarking","ai evaluation","DqEdE-KfCVokeZ_dDpJUjEgDp0i4PzcbAYQtN1PjJJA",{"id":303,"title":304,"author":8,"category":305,"date":306,"description":307,"extension":12,"featured":13,"image":308,"meta":309,"metaDescription":310,"metaTitle":311,"navigation":18,"path":312,"seo":313,"stem":314,"tags":315,"__hash__":320,"__readTime":151},"blog\u002Fblog\u002Fblits-ai-bug-hunt-game-2026-results.md","What We Learned from Turning Bug Hunting into a Game at Blits.ai","Product & Culture","2026-01-08","In the first week of January we turned quality assurance into a Bug Hunt Game. This is what happened when we put Hunters, Fixers and a Judge on a shared leaderboard powered by Azure DevOps.","\u002Fimages\u002Fblits-bounty-bug-game.png",{},"In January we turned quality assurance into a Bug Hunt Game, with Hunters, Fixers and a Judge on one Azure DevOps leaderboard. Here is what we learned.","Bug Hunting as a Game: What We Learned","\u002Fblog\u002Fblits-ai-bug-hunt-game-2026-results",{"title":304,"description":307},"blog\u002Fblits-ai-bug-hunt-game-2026-results",[316,317,318,319],"quality","devops","culture","experimentation","mWTOkdFZW8DPX4ik3aCDl_en1vapZ9YS-nCMmu1FFQU",{"id":322,"title":323,"author":275,"category":61,"date":324,"description":325,"extension":12,"featured":13,"image":326,"meta":327,"metaDescription":34,"metaTitle":328,"navigation":13,"path":329,"seo":330,"stem":331,"tags":332,"__hash__":333,"__readTime":334},"blog\u002Fblog\u002Fblits-ai-joins-llms4eu-project.md","Blits.ai joins the EU-funded LLMs4EU project","2026-01-05","Blits.ai joins LLMs4EU. Who builds large language models, on what data, and under which European values and regulations, and why that matters now.","\u002Fblog\u002Fllms4eu-blog-image.png",{},"LLMs4EU: Blits.ai Joins the European LLM Effort","\u002Fblog\u002Fblits-ai-joins-llms4eu-project",{"title":323,"description":325},"blog\u002Fblits-ai-joins-llms4eu-project",[],"M97zWsJIDe49eUcICylhQBw6aT5EKnTNuNWbZw178Pc","4 min read",{"id":336,"title":337,"author":8,"category":28,"date":338,"description":339,"extension":12,"featured":13,"image":340,"meta":341,"metaDescription":342,"metaTitle":343,"navigation":18,"path":344,"seo":345,"stem":346,"tags":34,"__hash__":347,"__readTime":71},"blog\u002Fblog\u002Fbusiness-predictions-chatgpt.md","Business predictions based on my years of experience with ChatGPT and its predecessors","2023-02-20","Over 5 years ago I started a chatbot (ad)venture called Blits.ai. My reason was simple: I believed the way we humans deal with data would turn into a conversation. Now that ChatGPT (generative AI) is all over the news and every LinkedIn feed, more and more people are reaching the same conclusion. Great! We're clearly not there yet, though.","\u002Fblog\u002F1675857460197.jpeg",{},"Five years after starting Blits.ai, my business predictions for ChatGPT and generative AI: where conversations with data are heading, and what is missing.","ChatGPT and Business: My Predictions","\u002Fblog\u002Fbusiness-predictions-chatgpt",{"title":337,"description":339},"blog\u002Fbusiness-predictions-chatgpt","26eZfqYo1gr5l1hXYbn1uJPqnvLUXu4T4vCAb5Gy39c",{"id":349,"title":350,"author":256,"body":351,"category":28,"date":697,"description":698,"extension":12,"featured":13,"image":699,"meta":700,"metaDescription":701,"metaTitle":702,"navigation":18,"path":703,"seo":704,"stem":705,"tags":706,"__hash__":709},"blog\u002Fblog\u002Fconversational-ai-design-why-bad-bots-are-a-design-problem.md","Conversational AI Design: Why Bad Bots Are a Design Problem, Not a Model Problem",{"type":352,"value":353,"toc":682},"minimark",[354,358,369,372,377,380,389,395,399,402,405,408,412,415,418,427,432,435,439,442,449,452,455,477,481,484,490,499,507,514,518,521,524,568,577,581,584,587,593,597,600,603,607,610,618,621,624,628,631,634,638,674],[355,356,357],"p",{},"It is ten past eleven at night and a customer types: \"My package never arrived.\" The assistant answers: \"Sorry, I didn't understand that.\" The customer tries again, gets a menu with three options that don't fit, and types the only thing left: \"Can I talk to a person?\" The assistant answers: \"Sorry, I didn't understand that.\"",[355,359,360,361,368],{},"The US Consumer Financial Protection Bureau has a name for this. It calls it a ",[362,363,367],"a",{"href":364,"rel":365},"https:\u002F\u002Fwww.consumerfinance.gov\u002Fdata-research\u002Fresearch-reports\u002Fchatbots-in-consumer-finance\u002Fchatbots-in-consumer-finance\u002F",[366],"noopener","doom loop",": automated responses that lead customers \"in continuous loops of repetitive, unhelpful jargon or legalese without an offramp to a human customer service representative.\" What strikes me every time I see one is that the model behind it is usually fine. The language understanding works. What failed is everything around it: nobody decided what the assistant was for, where it should stop, or what it should do when it got lost.",[355,370,371],{},"I'm a technical project manager at Blits.ai. Most of my week goes into building and testing conversational AI for banks and large enterprises, in English, Arabic dialects and Turkish, across chat and voice. This article collects what we have learned about designing conversations that hold up with real customers, organised around the questions teams ask us most.",[373,374,376],"h2",{"id":375},"what-is-conversational-ai-design","What is conversational AI design?",[355,378,379],{},"Conversational AI design is the practice of shaping how chatbots, voice agents and AI assistants understand people, guide a conversation, respond in the right tone and recover when something goes wrong. It covers goals, boundaries, wording, fallback paths and the handoff to a human. It is not the same as writing a prompt.",[355,381,382,383,388],{},"The distinction matters because customers judge the experience, not the model. Gartner found that ",[362,384,387],{"href":385,"rel":386},"https:\u002F\u002Fwww.gartner.com\u002Fen\u002Fnewsroom\u002Fpress-releases\u002F2024-07-09-gartner-survey-finds-64-percent-of-customers-would-prefer-that-companies-didnt-use-ai-for-customer-service",[366],"64 percent of customers would prefer companies didn't use AI in customer service",", and their biggest worry was that it would become harder to reach a person. Customers are reacting to the design they meet, whatever model sits behind it.",[390,391,392],"blockquote",{},[355,393,394],{},"A customer who cannot reach a person will blame the AI, whatever the model is.",[373,396,398],{"id":397},"where-should-conversation-design-start","Where should conversation design start?",[355,400,401],{},"Start with the job the customer is trying to get done, not with what the AI can do. \"Reschedule a delivery\", \"block my card\" and \"check my claim\" are jobs. \"Our bot can answer questions\" is not.",[355,403,404],{},"For every priority job, we map the conversation from the customer's side before anyone writes a prompt. The map starts with the goal, meaning the outcome the person wants and not only the topic they mention, and the information the assistant must collect to complete or route the request. It then describes the happy path, the shortest version of the conversation when everything is clear, and the unhappy paths: missing details, vague answers, emotional language, requests outside scope and system errors. Decision points mark where the assistant must choose between answering, asking, escalating or stopping. The last entry is the risk level: is this action trivial, sensitive, regulated, financial or impossible to undo?",[355,406,407],{},"The risk level is where banking teaches you humility. \"What is my balance?\" and \"block my card\" can arrive in the same chat, a minute apart, from the same anxious customer. They need completely different designs. The map forces product, compliance, operations and engineering to agree on that before the assistant goes live, instead of after the first complaint.",[373,409,411],{"id":410},"what-are-the-layers-of-a-good-conversational-interface","What are the layers of a good conversational interface?",[355,413,414],{},"Every conversational interface has three layers: understanding what the person means, steering the dialogue forward, and shaping the response. If any one of them fails, the whole experience feels broken, however powerful the model is.",[355,416,417],{},"Understanding means recognising intent, not matching keywords. \"I never got it\", \"my package is missing\" and \"it says delivered but it isn't here\" are the same problem. A good system uses context, account data and earlier turns to see that, and knows when it is not confident enough to act.",[355,419,420,421,426],{},"Understanding also means being grounded in approved sources: product documentation, policies, account data. An assistant without grounding sounds fluent while being wrong, and the company owns every word. When Air Canada's chatbot ",[362,422,425],{"href":423,"rel":424},"https:\u002F\u002Fwww.cbc.ca\u002Fnews\u002Fcanada\u002Fbritish-columbia\u002Fair-canada-chatbot-lawsuit-1.7116416",[366],"gave a passenger the wrong bereavement fare policy",", the airline argued that the bot was \"a separate legal entity that is responsible for its own actions.\" The tribunal disagreed and held Air Canada liable.",[390,428,429],{},[355,430,431],{},"Your chatbot is your company talking. Whatever it says, you said.",[355,433,434],{},"Dialogue decides how the conversation moves: when to let the person explain freely and when to follow a fixed sequence. Response covers what the assistant says and how: tone, length, structure and clarity, in chat and in voice.",[373,436,438],{"id":437},"when-should-the-ai-improvise-and-when-should-it-follow-a-script","When should the AI improvise, and when should it follow a script?",[355,440,441],{},"Let the AI reason freely when the customer is exploring or describing a problem. Make it propose and confirm before it changes anything. And make it follow a controlled flow for identity checks, payments, consent and anything regulated or impossible to undo.",[355,443,444],{},[445,446],"img",{"alt":447,"src":448},"Where the AI may improvise, and where it may not","\u002Fblog\u002Fblits-conversation-design-boundaries.png",[355,450,451],{},"We call these decision boundaries, and we draw them per task before writing a single prompt. The same assistant can move through all three zones in one conversation. Someone describes a problem in their own words (flexible), agrees to move a delivery (confirm first) and then asks for a refund to a new account (scripted). The design decides when the assistant switches, not the model's mood that day.",[355,453,454],{},"For the middle zone we use a simple pattern that customers find natural:",[456,457,458,462,465,468,471,474],"ol",{},[459,460,461],"li",{},"Recognise the intent: \"It sounds like you want to reschedule your delivery.\"",[459,463,464],{},"Confirm when unsure: \"Is that right?\"",[459,466,467],{},"Ask only for what's missing: \"Which day works better for you?\"",[459,469,470],{},"Explain the action: \"I can move it to Thursday afternoon.\"",[459,472,473],{},"Confirm before committing: \"Shall I make that change?\"",[459,475,476],{},"Close with the result: \"Done. Your delivery is now scheduled for Thursday afternoon.\"",[373,478,480],{"id":479},"how-should-fallback-and-human-handoff-work","How should fallback and human handoff work?",[355,482,483],{},"A fallback should always give the customer a way forward, and a handoff should pass the full context to the human so nobody has to start over. \"I didn't understand\" on its own gives the customer nothing to do next.",[355,485,486],{},[445,487],{"alt":488,"src":489},"The fallback ladder","\u002Fblog\u002Fblits-conversation-design-fallback.png",[355,491,492,493,498],{},"Nielsen Norman Group's research on chatbots said it plainly years ago: ",[362,494,497],{"href":495,"rel":496},"https:\u002F\u002Fwww.nngroup.com\u002Farticles\u002Fchatbots\u002F",[366],"be honest about not understanding, and offer an escape hatch"," in the form of a real human, a phone number or another channel. We design fallback as a ladder: clarify, suggest what the assistant can do, ask the customer to rephrase, stop safely at a risk boundary, and hand over. When the customer repeats themselves, uses urgent language or reaches a risk boundary, we skip straight to the top.",[355,500,501,502],{},"Klarna learned the cost of getting this wrong in public. After a year of promoting its AI assistant, its chief executive told Bloomberg in May 2025 that the company was hiring human agents again, and that it is critical ",[362,503,506],{"href":504,"rel":505},"https:\u002F\u002Fwww.entrepreneur.com\u002Fbusiness-news\u002Fklarna-ceo-reverses-course-by-hiring-more-humans-not-ai\u002F491396",[366],"\"that you are clear to your customer that there will always be a human if you want.\"",[355,508,509,510,513],{},"The handoff itself is where many designs fail. The human agent should receive the intent, a summary of the conversation, the details already collected, the steps already tried and any relevant system status. If the customer has to repeat everything, the AI has saved almost nothing. And for requests that run over days, the handoff is not the end of the conversation but one step in a case, which is why ",[362,511,512],{"href":48},"a conversation is not a case",".",[373,515,517],{"id":516},"what-should-the-opening-message-say","What should the opening message say?",[355,519,520],{},"The opening message should tell the customer they are talking to an AI assistant, what it can help with, and how to start. Short, honest and concrete beats warm and vague.",[355,522,523],{},"\"Hi, how can I help?\" invites anything, including everything the assistant cannot do. Compare:",[525,526,527,540],"table",{},[528,529,530],"thead",{},[531,532,533,537],"tr",{},[534,535,536],"th",{},"Use case",[534,538,539],{},"Opening line",[541,542,543,552,560],"tbody",{},[531,544,545,549],{},[546,547,548],"td",{},"Customer support",[546,550,551],{},"\"Hi, I'm the virtual assistant. I can track orders, start returns, change delivery details or connect you with a colleague.\"",[531,553,554,557],{},[546,555,556],{},"Appointments",[546,558,559],{},"\"I can book, move or cancel an appointment. What would you like to do?\"",[531,561,562,565],{},[546,563,564],{},"Troubleshooting",[546,566,567],{},"\"Describe what's going wrong and I'll suggest the next step. If we can't fix it here, I'll route you to the right team.\"",[355,569,570,571,576],{},"Saying it is AI is no longer just good manners. Article 50 of the EU AI Act requires that people are ",[362,572,575],{"href":573,"rel":574},"https:\u002F\u002Fartificialintelligenceact.eu\u002Farticle\u002F50\u002F",[366],"informed that they are interacting with an AI system",", and that obligation applies from 2 August 2026. The Nielsen Norman Group advice from the same research still holds as well: clearly tell people what tasks the bot can do, and don't create false expectations.",[373,578,580],{"id":579},"how-should-tone-and-persona-change-with-the-situation","How should tone and persona change with the situation?",[355,582,583],{},"Persona sets expectations about how the assistant communicates. It is not a quirky character. The right tone depends on the task and on the customer's state: calm and precise for billing, encouraging for troubleshooting, careful and never joking for health, money or security.",[355,585,586],{},"A useful persona guide describes how the assistant sounds when things go well, and how it sounds when the customer is confused, angry, worried or in a hurry. It also lists what the assistant never does: exaggerated enthusiasm, false certainty, humour in sensitive moments, or claims to feelings it doesn't have.",[355,588,589,590,513],{},"Language adds a layer most guides skip. In Arabic speaking markets, a perfectly correct answer in formal Modern Standard Arabic can still feel wrong to a customer who expects the way people in their own city speak. Register and dialect are part of the persona, and we test them with native speakers, not only with automatic scores. We wrote more about that in ",[362,591,592],{"href":80},"the voice quality gap",[373,594,596],{"id":595},"how-is-designing-for-voice-different-from-chat","How is designing for voice different from chat?",[355,598,599],{},"Chat can show options, links and summaries that the customer can read back. Voice cannot, so voice needs shorter turns, fewer choices and confirmation of critical details one at a time.",[355,601,602],{},"A voice agent might say: \"I found two options: Thursday morning or Friday afternoon. Which do you prefer?\" The same assistant in chat can show both as buttons. Long numbers are the classic voice trap: an account number read back in one breath is impossible to check. Read it in groups, and confirm it.",[373,604,606],{"id":605},"how-do-you-know-if-your-conversation-design-works","How do you know if your conversation design works?",[355,608,609],{},"Measure whether customers get their task done, not only whether they avoided a human. Containment is a useful number, but on its own it rewards exactly the doom loop this article started with.",[355,611,612,613,617],{},"Across the public deployments in our ",[362,614,616],{"href":615},"\u002Fai-use-cases\u002Fkpis\u002Fcontainment-rate","AI Use Case Library",", the median reported containment rate is 61.5 percent across 28 organisations, and 16 of those 28 values come from vendors. It is a fair benchmark, but it tells you nothing about the customers who gave up halfway.",[355,619,620],{},"The metrics we watch alongside containment: task completion, first contact resolution, repeat contacts within a few days, fallback rate, escalation quality, abandonment points and how often customers have to correct the assistant. And we read transcripts. The moments where people rephrase, repeat themselves or go quiet show design problems that no dashboard reveals.",[355,622,623],{},"Testing needs messy input: typos, slang, vague requests, angry messages, contradictory details. And it needs repeating whenever a prompt, policy, model or knowledge source changes, because conversational systems drift.",[373,625,627],{"id":626},"who-owns-the-conversation-once-it-is-live","Who owns the conversation once it is live?",[355,629,630],{},"Someone has to own every flow, prompt, knowledge source and metric, or the assistant slowly starts contradicting itself. Good governance keeps it coherent as products, policies and customers change.",[355,632,633],{},"That means named owners for flows and prompts, for the knowledge the assistant draws on, for compliance review and for quality assurance, with version control for all of it. It also means accessibility from the start: plain language, screen reader support, no meaning carried by colour alone, and room for people who need more time or simpler words. Designing for them makes the assistant better for everyone.",[373,635,637],{"id":636},"what-belongs-on-a-conversation-design-checklist","What belongs on a conversation design checklist?",[639,640,641,644,647,650,653,656,659,662,665,668,671],"ul",{},[459,642,643],{},"Start from the customer's task, not from the technology.",[459,645,646],{},"Tell customers they are talking to AI, and what it can do.",[459,648,649],{},"Ask one clear question at a time.",[459,651,652],{},"Confirm before anything sensitive or irreversible.",[459,654,655],{},"Ground every answer in approved sources.",[459,657,658],{},"Draw the boundaries: flexible, confirm first, scripted.",[459,660,661],{},"Design fallback as a ladder that ends with a human.",[459,663,664],{},"Pass the full context in every handoff.",[459,666,667],{},"Match tone, register and dialect to the customer and the risk.",[459,669,670],{},"Measure task completion and read transcripts, not only containment.",[459,672,673],{},"Retest whenever prompts, policies, models or knowledge sources change.",[355,675,676,677,513],{},"None of this makes a bot sound cleverer. It makes it useful, which is all the customer at ten past eleven wanted in the first place. If you are designing or rescuing an assistant and would like a second pair of eyes on the conversation, you can ",[362,678,681],{"href":679,"rel":680},"https:\u002F\u002Fwww.blits.ai\u002Fcontact-us",[366],"review your flows with us",{"title":683,"searchDepth":684,"depth":684,"links":685},"",2,[686,687,688,689,690,691,692,693,694,695,696],{"id":375,"depth":684,"text":376},{"id":397,"depth":684,"text":398},{"id":410,"depth":684,"text":411},{"id":437,"depth":684,"text":438},{"id":479,"depth":684,"text":480},{"id":516,"depth":684,"text":517},{"id":579,"depth":684,"text":580},{"id":595,"depth":684,"text":596},{"id":605,"depth":684,"text":606},{"id":626,"depth":684,"text":627},{"id":636,"depth":684,"text":637},"2026-10-21","Most chatbots that frustrate people run on perfectly capable models. What fails is the design: unclear goals, no boundaries, fallbacks that go nowhere and handoffs that lose the context. A practical guide in questions and answers, from the team that builds and tests these systems every day.","\u002Fblog\u002Fblits-conversation-design-hero.png",{},"A practical guide to conversational AI design: user goals, decision boundaries, fallback and human handoff, opening messages, tone, voice and measuring quality.","Conversational AI Design: A Practical Guide","\u002Fblog\u002Fconversational-ai-design-why-bad-bots-are-a-design-problem",{"title":350,"description":698},"blog\u002Fconversational-ai-design-why-bad-bots-are-a-design-problem",[216,707,708,268,199],"conversation design","chatbot ux","80xT4JDMkcCexuKiC7QG6RvgmV58yMZYsNutW_u-wiU",{"id":711,"title":712,"author":275,"category":28,"date":713,"description":714,"extension":12,"featured":13,"image":715,"meta":716,"metaDescription":34,"metaTitle":717,"navigation":18,"path":718,"seo":719,"stem":720,"tags":721,"__hash__":723,"__readTime":334},"blog\u002Fblog\u002Feu-ai-act-2026-enterprise-readiness-checklist.md","EU AI Act 2026: The Enterprise AI Readiness Checklist Before August","2026-03-11","Most teams treat EU AI Act compliance as a legal side project. A checklist of what enterprises should implement now to avoid disruption and rework.","\u002Fimages\u002Fblog-eu-ai-act.png",{},"EU AI Act 2026: Enterprise Readiness Checklist","\u002Fblog\u002Feu-ai-act-2026-enterprise-readiness-checklist",{"title":712,"description":714},"blog\u002Feu-ai-act-2026-enterprise-readiness-checklist",[722,53,118,270],"eu ai act","rluJvoZAzD04WieClVwCsoj6r18ndtcHInhwIzsy0v4",{"id":725,"title":726,"author":8,"category":28,"date":727,"description":728,"extension":12,"featured":13,"image":729,"meta":730,"metaDescription":34,"metaTitle":731,"navigation":18,"path":732,"seo":733,"stem":734,"tags":34,"__hash__":735,"__readTime":201},"blog\u002Fblog\u002Ffeeding-llms-without-leaking-secrets.md","Feeding LLMs Without Leaking Secrets: A Guide for Companies On How To Add Your Company Data","2025-05-19","A plain guide for business teams on how to add your company data to an LLM safely, without leaking secrets, and the concepts behind it.","\u002Fblog\u002F1747588091563.png",{},"Feeding LLMs Without Leaking Company Secrets","\u002Fblog\u002Ffeeding-llms-without-leaking-secrets",{"title":726,"description":728},"blog\u002Ffeeding-llms-without-leaking-secrets","8IxE-B9QPvTgygMq-My29zKji23d5TeHn16xmfgIgYY",{"id":737,"title":738,"author":256,"category":739,"date":740,"description":741,"extension":12,"featured":13,"image":742,"meta":743,"metaDescription":34,"metaTitle":744,"navigation":18,"path":745,"seo":746,"stem":747,"tags":748,"__hash__":750,"__readTime":252},"blog\u002Fblog\u002Ffrom-johannesburg-five-signals-agentic-ai-enterprise-banking.md","From Johannesburg: Five Signals on Where Agentic AI Meets Enterprise Banking","Events","2026-07-15","Five takeaways from Mastercard's summit in Johannesburg: the technology is no longer the bottleneck, governance, trust and context are.","\u002Fblog\u002Ffive-signals-mastercard.png",{},"Five Signals on Agentic AI in Enterprise Banking","\u002Fblog\u002Ffrom-johannesburg-five-signals-agentic-ai-enterprise-banking",{"title":738,"description":741},"blog\u002Ffrom-johannesburg-five-signals-agentic-ai-enterprise-banking",[52,749,216,248,300],"enterprise banking","b_mlFIOndF2DPQRqZDiXZNvaVGRyPPI3P5jsTOTRhQY",{"id":752,"title":753,"author":136,"category":739,"date":754,"description":755,"extension":12,"featured":18,"image":756,"meta":757,"metaDescription":34,"metaTitle":758,"navigation":18,"path":759,"seo":760,"stem":761,"tags":762,"__hash__":763,"__readTime":151},"blog\u002Fblog\u002Ffrom-paris-to-rio-signals-and-where-enterprise-ai-is-heading.md","From Paris to Rio: Production Over Pilots, Substance Over Spectacle","2026-07-10","Signals from VivaTech and Web Summit Rio: production over pilots, substance over spectacle, and why your operating framework beats any one model.","\u002Fblog\u002Ffrom-paris-to-rio.png",{},"From Paris to Rio: Where Enterprise AI Is Heading","\u002Fblog\u002Ffrom-paris-to-rio-signals-and-where-enterprise-ai-is-heading",{"title":753,"description":755},"blog\u002Ffrom-paris-to-rio-signals-and-where-enterprise-ai-is-heading",[53,52,216,54,118],"6DzNE9FCbGcnNxwm9Du9-SwxJ5yh585XO0lPY3tLxng",{"id":765,"title":766,"author":275,"category":28,"date":767,"description":768,"extension":12,"featured":13,"image":769,"meta":770,"metaDescription":34,"metaTitle":771,"navigation":18,"path":772,"seo":773,"stem":774,"tags":775,"__hash__":778,"__readTime":779},"blog\u002Fblog\u002Ffrom-single-agent-to-multi-agent-systems-when-to-split-roles.md","Most Multi Agent Systems Are a Solution Looking for a Problem","2026-03-19","Start with one agent, and reach for multi agent only for parallel, read heavy work. The case against splitting too early, and how to do it right.","\u002Fimages\u002Fblog-multi-agent.png",{},"When to Split Into Multi Agent Systems","\u002Fblog\u002Ffrom-single-agent-to-multi-agent-systems-when-to-split-roles",{"title":766,"description":768},"blog\u002Ffrom-single-agent-to-multi-agent-systems-when-to-split-roles",[52,776,53,149,777],"multi agent systems","llm","jh1LuV8_t23zMxcRv4qwpkRBlmGeW4eHuFvagOyENQM","10 min read",{"id":781,"title":782,"author":8,"category":9,"date":783,"description":784,"extension":12,"featured":13,"image":785,"meta":786,"metaDescription":34,"metaTitle":787,"navigation":18,"path":788,"seo":789,"stem":792,"tags":793,"__hash__":794,"__readTime":71},"blog\u002Fblog\u002Ffrom-talking-to-doing-six-months-of-blits-ai-updates.md","Blits.ai Platform Update: Five Big Additions from the Last Six Months","2026-07-20","A plain roundup of five things we shipped this year: emotional voice, interactive cards, agentic tasks, in chat payments and response feedback.","\u002Fblog\u002Fblits-login-screen.png",{},"Blits.ai Platform Update: Six Months of Additions","\u002Fblog\u002Ffrom-talking-to-doing-six-months-of-blits-ai-updates",{"title":790,"description":791},"Product Update: Five Big Additions from the Last Six Months","A plain language roundup of the five biggest things we shipped this year (emotional voice, rich interactive cards, agentic tasks, in chat payments and response feedback) and what each one does for the person on the other end of the conversation.","blog\u002Ffrom-talking-to-doing-six-months-of-blits-ai-updates",[165,52,216,268,248],"yojUU4dXMnn1LM4XJdH8rXr-2_-CI5k9oYKfDLjNOwo",{"id":796,"title":797,"author":275,"category":28,"date":798,"description":799,"extension":12,"featured":13,"image":800,"meta":801,"metaDescription":34,"metaTitle":802,"navigation":18,"path":803,"seo":804,"stem":805,"tags":34,"__hash__":806,"__readTime":285},"blog\u002Fblog\u002Fgpt-4-in-chatbots.md","GPT-4 in Chatbots: Blits.ai Breaks New Ground in Conversational AI","2023-07-14","Blits.ai now supports GPT-4, bringing new capabilities to chatbot development and deployment. What it means for conversational AI.","\u002Fblog\u002Fgpt-4.webp",{},"GPT-4 Chatbots: New Ground in Conversational AI","\u002Fblog\u002Fgpt-4-in-chatbots",{"title":797,"description":799},"blog\u002Fgpt-4-in-chatbots","G00HNe_XJ2lwAPzMes8evMiwWORzWzdgIFCzLrsptWU",{"id":808,"title":809,"author":275,"category":28,"date":810,"description":811,"extension":12,"featured":13,"image":812,"meta":813,"metaDescription":814,"metaTitle":815,"navigation":18,"path":816,"seo":817,"stem":818,"tags":34,"__hash__":819,"__readTime":285},"blog\u002Fblog\u002Fhow-practical-is-gpt-3-for-conversational-ai-chatbots.md","How practical is GPT-3 for Conversational AI chatbots?","2022-01-25","GPT-3 has changed what people expect from conversational AI. Ask it almost anything and you get a fluent, often surprisingly good answer. Whether that makes it practical for a chatbot or voicebot that talks to your customers is a different question, and that is what this article is about.","\u002Fblog\u002FGPT-3.jpg",{},"GPT-3 gives fluent, often surprisingly good answers. Whether that makes it practical for a chatbot that talks to your customers is another question.","How Practical Is GPT-3 for Chatbots?","\u002Fblog\u002Fhow-practical-is-gpt-3-for-conversational-ai-chatbots",{"title":809,"description":811},"blog\u002Fhow-practical-is-gpt-3-for-conversational-ai-chatbots","5bcufuhpdudmHxSJDIeisrRbO9tOk1zdavIsJZcNhBY",{"id":821,"title":822,"author":8,"category":28,"date":823,"description":824,"extension":12,"featured":13,"image":825,"meta":826,"metaDescription":827,"metaTitle":828,"navigation":18,"path":829,"seo":830,"stem":831,"tags":832,"__hash__":835,"__readTime":252},"blog\u002Fblog\u002Fhow-to-build-an-ai-control-tower-for-agentic-operations.md","The AI Control Tower: Air Traffic Control for Agents in Production","2026-03-23","A dashboard tells you the plane landed. A control tower keeps two aircraft off the same runway in real time, with the authority to divert one. What that distinction means for running AI agents in production, and why an HTTP 200 will lie to you.","\u002Fimages\u002Fblog-control-tower.png",{},"A dashboard tells you the plane landed. A control tower keeps two aircraft off one runway. What that means for running AI agents in production.","How to Build an AI Control Tower for Agents","\u002Fblog\u002Fhow-to-build-an-ai-control-tower-for-agentic-operations",{"title":822,"description":824},"blog\u002Fhow-to-build-an-ai-control-tower-for-agentic-operations",[833,52,118,834,53],"ai observability","guardrails","tekVT7bbrPjtuBBrGEuTtVN495S40tyKTdMJ6WTqSmQ",{"id":837,"title":838,"author":275,"category":9,"date":839,"description":840,"extension":12,"featured":13,"image":841,"meta":842,"metaDescription":843,"metaTitle":844,"navigation":18,"path":845,"seo":846,"stem":847,"tags":848,"__hash__":852,"__readTime":39},"blog\u002Fblog\u002Fintroducing-blits-ai-financial-wellbeing.md","Introducing Blits.ai Financial Wellbeing: Guidance, Never Financial Advice","2026-08-29","Our newest product turns a bank's own transaction data into a wellbeing score, plain language insight and gentle nudges toward healthier money habits. It runs inside the bank, it measures its own impact, and it will never give financial advice. Here is how it works.","\u002Fblog\u002Fblits-fw-dashboard.png",{},"Our newest product turns a bank's transaction data into a wellbeing score, plain language insight and gentle nudges. It runs inside the bank, never as advice.","Blits.ai Financial Wellbeing: Guidance, Not Advice","\u002Fblog\u002Fintroducing-blits-ai-financial-wellbeing",{"title":838,"description":840},"blog\u002Fintroducing-blits-ai-financial-wellbeing",[849,749,834,850,851],"financial wellbeing","product","nudges","fr19gXuC3QnydGpIKcJpc9c9ZjA6F-HESxQnaqK8dfY",{"id":854,"title":855,"author":256,"category":28,"date":856,"description":857,"extension":12,"featured":13,"image":858,"meta":859,"metaDescription":860,"metaTitle":861,"navigation":18,"path":862,"seo":863,"stem":864,"tags":865,"__hash__":870,"__readTime":39},"blog\u002Fblog\u002Fknowledge-graphs-vs-vector-search.md","Knowledge Graphs vs Vector Search: What Enterprise AI Actually Needs","2026-11-24","Vector search finds passages that mean the same thing. A knowledge graph knows how things are connected. Most enterprise AI needs the first, some needs the second, and a surprising number of questions need neither. A practical guide to choosing, from someone who builds both.","\u002Fblog\u002Fblits-knowledge-graph-hero.png",{},"Knowledge graph or vector search? What each one is good at, where each breaks, and how to decide which questions your enterprise AI needs to answer.","Knowledge Graphs vs Vector Search for Enterprise AI","\u002Fblog\u002Fknowledge-graphs-vs-vector-search",{"title":855,"description":857},"blog\u002Fknowledge-graphs-vs-vector-search",[866,867,868,869,53],"knowledge graph","vector search","rag","retrieval","IPa50HdvBjLxCAI5Q_SAlhu7UaoQewgBoWLBooU0CkQ",{"id":872,"title":873,"author":275,"category":28,"date":874,"description":875,"extension":12,"featured":13,"image":876,"meta":877,"metaDescription":34,"metaTitle":878,"navigation":18,"path":879,"seo":880,"stem":881,"tags":882,"__hash__":886,"__readTime":252},"blog\u002Fblog\u002Fmcp-security-in-production-auth-and-least-privilege.md","MCP Security: The Threat Model Changes the Day After You Ship","2026-03-15","Most MCP breaches needed no exploit code; the servers behaved as specified. What goes wrong in production, and why the new OAuth spec is half a fix.","\u002Fimages\u002Fblog-mcp-security.png",{},"MCP Security in Production: Auth and Privilege","\u002Fblog\u002Fmcp-security-in-production-auth-and-least-privilege",{"title":873,"description":875},"blog\u002Fmcp-security-in-production-auth-and-least-privilege",[883,884,885,53,52],"mcp","ai security","prompt injection","Hz41mPrcsT9FC-R_MkrwnwIFerSFFo6bu9tJw9xeKTA",{"id":888,"title":889,"author":8,"category":28,"date":890,"description":891,"extension":12,"featured":13,"image":892,"meta":893,"metaDescription":34,"metaTitle":894,"navigation":18,"path":895,"seo":896,"stem":897,"tags":898,"__hash__":901,"__readTime":334},"blog\u002Fblog\u002Fmcp-servers-and-tools.md","MCP Servers and Tools: How LLMs Connect to the Real World","2026-01-27","MCP servers standardize how LLMs find and use tools. What they are, why they matter, how we use them at Blits.ai, and where this is going next.","\u002Fimages\u002Fmcp-server-header.webp",{},"MCP Servers and Tools: How LLMs Reach the World","\u002Fblog\u002Fmcp-servers-and-tools",{"title":889,"description":891},"blog\u002Fmcp-servers-and-tools",[777,899,883,900],"tools","agents","H-rIH_hhI4AEIuhMm5XBRn8GqeNa7Ie_6MmiMlzHllw",{"id":903,"title":904,"author":275,"category":739,"date":905,"description":906,"extension":12,"featured":13,"image":907,"meta":908,"metaDescription":909,"metaTitle":910,"navigation":18,"path":911,"seo":912,"stem":913,"tags":914,"__hash__":918,"__readTime":71},"blog\u002Fblog\u002Fmeasurable-beats-impressive-blits-ai-at-ai-everything-abu-dhabi.md","Measurable Beats Impressive: 75 Conversations at Ai Everything Abu Dhabi","2026-10-15","Len and I spent two days on our stand at Ai Everything Abu Dhabi and had more than 75 conversations with banks, government entities, telcos, investors and fellow founders. Five themes kept coming back, and together they point one way: from impressive demos to measurable results in production.","\u002Fblog\u002Fblits-ai-everything-abu-dhabi-recap-hero.png",{},"What more than 75 conversations at Ai Everything Abu Dhabi taught us about enterprise AI in the Gulf: production, sovereignty, agents, language and trust.","75 Conversations at Ai Everything Abu Dhabi 2026","\u002Fblog\u002Fmeasurable-beats-impressive-blits-ai-at-ai-everything-abu-dhabi",{"title":904,"description":906},"blog\u002Fmeasurable-beats-impressive-blits-ai-at-ai-everything-abu-dhabi",[915,916,52,917,53],"events","abu dhabi","financial services","SAo3YW1KhOwDpp96BK4_MIozyIYm8kJouTj-uCO9Bog",{"id":920,"title":921,"author":8,"category":28,"date":922,"description":923,"extension":12,"featured":13,"image":924,"meta":925,"metaDescription":34,"metaTitle":34,"navigation":18,"path":926,"seo":927,"stem":928,"tags":929,"__hash__":933,"__readTime":39},"blog\u002Fblog\u002Fmeasure-ai-performance-and-set-the-right-kpis.md","Measure AI Performance and Set the Right KPIs","2026-03-07","AI performance means more than accuracy. These are the production KPIs that matter (latency, P99, TTFT and error rates) and how to measure them.","\u002Fimages\u002Fblits-performance.jpg",{},"\u002Fblog\u002Fmeasure-ai-performance-and-set-the-right-kpis",{"title":921,"description":923},"blog\u002Fmeasure-ai-performance-and-set-the-right-kpis",[930,931,932,53],"ai kpi","ai performance","observability","7izChjS6RtzVMstKY8mFca2uii-3N31S4GP1zVQOxX0",{"id":935,"title":936,"author":8,"category":28,"date":937,"description":938,"extension":12,"featured":13,"image":939,"meta":940,"metaDescription":34,"metaTitle":34,"navigation":18,"path":941,"seo":942,"stem":943,"tags":944,"__hash__":946,"__readTime":103},"blog\u002Fblog\u002Fnobody-budgets-in-tokens.md","Nobody Budgets in Tokens","2026-08-12","Per seat licensing carried software for a decade, but agents don't sit in seats and token pricing breaks the moment AI succeeds. Where AI pricing lands.","\u002Fblog\u002Fblits-pricing-eras-hero.png",{},"\u002Fblog\u002Fnobody-budgets-in-tokens",{"title":936,"description":938},"blog\u002Fnobody-budgets-in-tokens",[945,53,52,54],"ai pricing","hkBVqOB8OWhD605LVQ0T2z7EUAGFurKvJc45xadSVxs",{"id":948,"title":949,"author":275,"category":28,"date":950,"description":951,"extension":12,"featured":13,"image":952,"meta":953,"metaDescription":34,"metaTitle":954,"navigation":18,"path":955,"seo":956,"stem":957,"tags":34,"__hash__":958,"__readTime":24},"blog\u002Fblog\u002Fopening-the-natural-language-understanding-nlu-blackbox.md","Opening the Natural Language Understanding (NLU) Blackbox","2020-06-23","NLU is often treated as a black box, but knowing how it works is key to building effective conversational AI. A look inside.","\u002Fblog\u002FMonitoring-300x234.png",{},"Opening the NLU Black Box","\u002Fblog\u002Fopening-the-natural-language-understanding-nlu-blackbox",{"title":949,"description":951},"blog\u002Fopening-the-natural-language-understanding-nlu-blackbox","qPChjo2cCcMnFSlIS91oCjCT84N-DppSPKyROYaYC2Q",{"id":960,"title":961,"author":256,"category":28,"date":962,"description":963,"extension":12,"featured":13,"image":964,"meta":965,"metaDescription":966,"metaTitle":967,"navigation":18,"path":968,"seo":969,"stem":970,"tags":971,"__hash__":972,"__readTime":151},"blog\u002Fblog\u002Fprompt-injection-enterprise-ai-agents.md","Prompt Injection: The Attack Every Enterprise AI Agent Will Face","2026-11-20","Prompt injection is the number one risk on the OWASP list for LLM applications, and security agencies say it may never be fully solved. For AI agents that read emails and documents and act in your systems, that changes how you design them. What prompt injection is, real incidents, and five layers that keep a successful attack from doing real harm.","\u002Fblog\u002Fblits-prompt-injection-hero.png",{},"What prompt injection is, why it is the top risk for LLM applications, why it cannot be fully solved, and five layers that limit the damage for AI agents.","Prompt Injection: The Attack Every AI Agent Will Face","\u002Fblog\u002Fprompt-injection-enterprise-ai-agents",{"title":961,"description":963},"blog\u002Fprompt-injection-enterprise-ai-agents",[885,884,52,118,56],"mBu03eY3ykp-MVizY4m6FATerfgXKC3E7Rka_o9M8vc",{"id":974,"title":975,"author":8,"category":28,"date":976,"description":977,"extension":12,"featured":13,"image":978,"meta":979,"metaDescription":34,"metaTitle":980,"navigation":18,"path":981,"seo":982,"stem":983,"tags":984,"__hash__":985,"__readTime":779},"blog\u002Fblog\u002Frag-in-2026-why-enterprise-pipelines-still-fail.md","The RAG That Passed Its Demo and Quietly Died in Production","2026-04-16","A knowledge assistant that aces its pilot can fail silently, recall sliding while dashboards stay green. Why good RAG rots after the demo.","\u002Fimages\u002Fblits-hallucination.jpg",{},"RAG in 2026: Why Enterprise Pipelines Still Fail","\u002Fblog\u002Frag-in-2026-why-enterprise-pipelines-still-fail",{"title":975,"description":977},"blog\u002Frag-in-2026-why-enterprise-pipelines-still-fail",[868,53,777,869,300],"30WucP_fwRnTHUJ8UHdi6VXrjrevV1S2Uhl6sQMEEMY",{"id":987,"title":988,"author":8,"category":28,"date":989,"description":990,"extension":12,"featured":13,"image":991,"meta":992,"metaDescription":993,"metaTitle":988,"navigation":18,"path":994,"seo":995,"stem":996,"tags":997,"__hash__":1001,"__readTime":39},"blog\u002Fblog\u002Fretail-ai-use-cases-that-survive-the-demo.md","Retail AI: The Use Cases That Survive Beyond the Demo","2026-11-27","Retail AI demos are all shopping assistants that talk like a friend with great taste. The retail AI that survives in production is mostly pricing, forecasting, fraud screening and product data. What 123 public deployments tell us, what Walmart, Amazon and IKEA actually report, and where the shopping assistant finally earns its place.","\u002Fblog\u002Fblits-retail-ai-hero.png",{},"Which retail AI use cases make it to production? What 123 public deployments, Walmart, Amazon, IKEA and Lowe's show, and why the boring ones pay first.","\u002Fblog\u002Fretail-ai-use-cases-that-survive-the-demo",{"title":988,"description":990},"blog\u002Fretail-ai-use-cases-that-survive-the-demo",[998,999,52,199,1000],"retail ai","ecommerce","ai use cases","-ud86QKo5OoTfyhYMSq4I40xq-UYZWhtOdJTUPcddHU",{"id":1003,"title":1004,"author":8,"category":28,"date":1005,"description":1006,"extension":12,"featured":13,"image":1007,"meta":1008,"metaDescription":34,"metaTitle":1009,"navigation":18,"path":1010,"seo":1011,"stem":1012,"tags":1013,"__hash__":1015,"__readTime":39},"blog\u002Fblog\u002Ftext-to-speech-engines-and-why-they-matter.md","Text-to-Speech Engines: The Voice Layer Every AI Product Needs","2026-03-03","Text to speech is core AI infrastructure. What TTS engines are, the model approaches, the providers that matter, and why speed and dialect decide.","\u002Fimages\u002Fblits-voice.jpg",{},"Text to Speech Engines: The Voice Layer for AI","\u002Fblog\u002Ftext-to-speech-engines-and-why-they-matter",{"title":1004,"description":1006},"blog\u002Ftext-to-speech-engines-and-why-they-matter",[298,268,1014],"ai infrastructure","Eu3je_dRENeJZno3LYJ_fO5Q8qAgtwBhXUoOKz3wGcA",{"id":1017,"title":1018,"author":8,"category":28,"date":1019,"description":1020,"extension":12,"featured":13,"image":1021,"meta":1022,"metaDescription":1023,"metaTitle":34,"navigation":18,"path":1024,"seo":1025,"stem":1026,"tags":1027,"__hash__":1029,"__readTime":201},"blog\u002Fblog\u002Fthe-ai-bank-runs-on-governance.md","The AI Bank Runs on Governance","2026-09-30","Our banking use case map grew from 21 to 81 jobs, so we turned it into a playbook. Along the way the story changed: the destination is the AI bank, where the work itself runs on governed AI, and what decides who gets there is governance far more than the models.","\u002Fblog\u002Fblits-ai-bank-eras-hero.png",{},"Our banking use case map grew from 21 to 81 jobs. The destination is the AI bank, and what decides who gets there is governance more than the models.","\u002Fblog\u002Fthe-ai-bank-runs-on-governance",{"title":1018,"description":1020},"blog\u002Fthe-ai-bank-runs-on-governance",[1028,118,749,52,270],"ai bank","ziFVXhwwYmus7n690c2_Rwb-4y8O4_M3dywY49MRwnc",{"id":1031,"title":1032,"author":8,"category":28,"date":1033,"description":1034,"extension":12,"featured":13,"image":1035,"meta":1036,"metaDescription":34,"metaTitle":34,"navigation":18,"path":1037,"seo":1038,"stem":1039,"tags":1040,"__hash__":1042,"__readTime":201},"blog\u002Fblog\u002Fthe-best-ai-model-doesnt-exist.md","The Best AI Model Doesn't Exist","2026-07-27","62 AI model versions got a retirement date this year. Why locking your business to one model is a hidden risk, and why staying model agnostic wins.","\u002Fblog\u002Fblits-model-agnostic-hub.png",{},"\u002Fblog\u002Fthe-best-ai-model-doesnt-exist",{"title":1032,"description":1034},"blog\u002Fthe-best-ai-model-doesnt-exist",[1041,53,777,299,54],"model agnostic","JqdV1Cu_gxRCBVp7H6x-wfyip2yVMkHRo7Tu0G_4gnM",{"id":1044,"title":1045,"author":8,"category":61,"date":1046,"description":1047,"extension":12,"featured":13,"image":1048,"meta":1049,"metaDescription":1050,"metaTitle":1051,"navigation":18,"path":1052,"seo":1053,"stem":1054,"tags":34,"__hash__":1055,"__readTime":24},"blog\u002Fblog\u002Fthe-best-way-for-chatbots-to-actually-understand-your-customers.md","Blits.ai: The Best Way for Chatbots to Actually Understand Your Customers","2021-02-11","We're introducing a new method at Blits.ai that changes how chatbots work under the hood. From the start our idea was simple: companies should be able to use the best performing engines on the market without having to pick one vendor and hope for the best. Right now we are two things at once, a middleware layer and a low code platform, and through it our customers get access to 40+ cognitive AI services.","\u002Fblog\u002F1611759270728.png",{},"A new Blits.ai method lets chatbots use the best AI engines on the market without betting on one vendor, through 40+ cognitive AI services.","How Chatbots Truly Understand Your Customers","\u002Fblog\u002Fthe-best-way-for-chatbots-to-actually-understand-your-customers",{"title":1045,"description":1047},"blog\u002Fthe-best-way-for-chatbots-to-actually-understand-your-customers","jlCmU9T6D-chVvdsZIMpRg5Gk4hbzvt7Mut-plLU6ZY",{"id":1057,"title":1058,"author":275,"category":28,"date":1059,"description":1060,"extension":12,"featured":13,"image":1061,"meta":1062,"metaDescription":1063,"metaTitle":34,"navigation":18,"path":1064,"seo":1065,"stem":1066,"tags":1067,"__hash__":1070,"__readTime":39},"blog\u002Fblog\u002Fthe-chargeback-arms-race.md","The Chargeback Arms Race","2026-10-03","Chargebacks are heading for 337 million a year, most of them aren't real fraud, and each one costs a multiple of its face value. A practical look at the six AI jobs hiding in one dispute, what each of them needs to be governed, and why agentic commerce will change the volume.","\u002Fblog\u002Fblits-chargeback-loop-hero.png",{},"Chargebacks head for 337 million a year and most are not fraud. The six AI jobs in one dispute, how to govern each, and what agentic commerce changes.","\u002Fblog\u002Fthe-chargeback-arms-race",{"title":1058,"description":1060},"blog\u002Fthe-chargeback-arms-race",[1068,250,52,53,1069],"chargebacks","fraud","HPFO_kwJvdXKmDRJf83bSyREvy4-rH815TsB1ButS64",{"id":1072,"title":1073,"author":136,"category":28,"date":1074,"description":1075,"extension":12,"featured":13,"image":1076,"meta":1077,"metaDescription":1078,"metaTitle":34,"navigation":18,"path":1079,"seo":1080,"stem":1081,"tags":1082,"__hash__":1085,"__readTime":201},"blog\u002Fblog\u002Fthe-company-that-calls-first.md","The Company That Calls First","2026-09-23","A brief landed on my desk asking for 600,000 outbound calls a month, by AI. The pattern is not a banking story: it applies to every enterprise with customers worth calling. Why human dialling died, what insurers, banks, lenders and software companies already run, and the anatomy of a governed outbound call.","\u002Fblog\u002Fblits-outbound-hero.png",{},"A brief asked for 600,000 outbound AI calls a month. Why human dialling died, what insurers, banks and lenders already run, and how a governed outbound call works.","\u002Fblog\u002Fthe-company-that-calls-first",{"title":1073,"description":1075},"blog\u002Fthe-company-that-calls-first",[1083,268,53,52,1084],"outbound ai","sales","lex9hp4VJze5EHsfvgwSNDTKzRgcx7xuHKtS3Co1hxc",{"id":1087,"title":1088,"author":275,"category":28,"date":1089,"description":1090,"extension":12,"featured":13,"image":1091,"meta":1092,"metaDescription":1093,"metaTitle":1094,"navigation":18,"path":1095,"seo":1096,"stem":1097,"tags":1098,"__hash__":1100,"__readTime":151},"blog\u002Fblog\u002Fthe-coordination-layer-what-to-build-when-models-converge.md","The Coordination Layer: What to Build When Models Converge","2026-10-20","Models are converging and getting cheaper to swap, so the durable engineering in enterprise AI is moving somewhere else: the layer that decides who may act, what is remembered, when a human takes over and what gets evidenced. A build guide for the part of your AI stack you'll still own in five years.","\u002Fblog\u002Fblits-coordination-layer-hero.png",{},"Models converge and get cheaper to swap, so lasting work moves to the layer that decides who acts, what is remembered and when a human takes over.","What to Build When AI Models Converge","\u002Fblog\u002Fthe-coordination-layer-what-to-build-when-models-converge",{"title":1088,"description":1090},"blog\u002Fthe-coordination-layer-what-to-build-when-models-converge",[149,52,53,118,1099],"integration","6bV3tAmavHYlr6iTqMhgdg-TOmMYZipHTMmtAENHZCE",{"id":1102,"title":1103,"author":8,"category":28,"date":1104,"description":1105,"extension":12,"featured":13,"image":978,"meta":1106,"metaDescription":1107,"metaTitle":1108,"navigation":18,"path":1109,"seo":1110,"stem":1111,"tags":1112,"__hash__":1115,"__readTime":201},"blog\u002Fblog\u002Fthe-danger-of-ai-hallucinations-and-how-businesses-should-handle-it.md","The Danger of AI Hallucinations and How Businesses Should Handle It","2026-02-25","AI hallucinations are a business risk as much as an error. What they are, why models make things up, how often it happens, real cases from courts and companies, practical mitigations, and what you gain by getting it right.",{},"What AI hallucinations are, why models make things up, how often they do, real cases from courts and companies, and how businesses reduce the risk.","AI Hallucinations and How Businesses Handle Them","\u002Fblog\u002Fthe-danger-of-ai-hallucinations-and-how-businesses-should-handle-it",{"title":1103,"description":1105},"blog\u002Fthe-danger-of-ai-hallucinations-and-how-businesses-should-handle-it",[1113,777,1114,53],"ai","risk management","hXB0xURMAestXPn4SvNI1RX6xuzd8zazx12PKSMkFJU",{"id":1117,"title":1118,"author":8,"category":28,"date":1119,"description":1120,"extension":12,"featured":13,"image":1121,"meta":1122,"metaDescription":34,"metaTitle":1123,"navigation":18,"path":1124,"seo":1125,"stem":1126,"tags":1127,"__hash__":1129,"__readTime":779},"blog\u002Fblog\u002Fthe-many-forms-of-rag-explained-without-the-jargon.md","The Many Forms of RAG, Explained Without the Jargon","2026-07-02","Hybrid search, reranking, ColBERT, GraphRAG, self-RAG: a plain language guide to what each RAG variant does and which ones you probably don't need.","\u002Fblog\u002FBlits-ai-RAG-used-to-be-one-thing.png",{},"The Many Forms of RAG, Without the Jargon","\u002Fblog\u002Fthe-many-forms-of-rag-explained-without-the-jargon",{"title":1118,"description":1120},"blog\u002Fthe-many-forms-of-rag-explained-without-the-jargon",[868,53,777,869,1128],"ai explained","I_4mb1aJRQPXhlA7DS18iNF9x69fHflEzyBjmTit6RE",{"id":1131,"title":1132,"author":256,"category":28,"date":1133,"description":1134,"extension":12,"featured":13,"image":1135,"meta":1136,"metaDescription":34,"metaTitle":1137,"navigation":18,"path":1138,"seo":1139,"stem":1140,"tags":1141,"__hash__":1143,"__readTime":1144},"blog\u002Fblog\u002Fthe-missing-text-layer-why-speech-to-speech-fails-enterprise.md","The Missing Text Layer: Why Speech-to-Speech Agents Fail Enterprise Requirements","2026-07-23","Native speech to speech demos beautifully, then fails enterprise needs. Why the text layer, where guardrails and audit live, still wins regulated voice.","\u002Fblog\u002Fblitsai-voice-isnt-safe.png",{},"Why Speech to Speech Voice Agents Fail Enterprise","\u002Fblog\u002Fthe-missing-text-layer-why-speech-to-speech-fails-enterprise",{"title":1132,"description":1134},"blog\u002Fthe-missing-text-layer-why-speech-to-speech-fails-enterprise",[268,1142,53,149,216],"speech to speech","_wRru7BtLUWa3kxmqXFqOH19LuZw0aj0P7uY7jSmSYc","15 min read",{"id":1146,"title":1147,"author":8,"category":28,"date":1148,"description":1149,"extension":12,"featured":13,"image":1150,"meta":1151,"metaDescription":1152,"metaTitle":1153,"navigation":18,"path":1154,"seo":1155,"stem":1156,"tags":1157,"__hash__":1158,"__readTime":201},"blog\u002Fblog\u002Fthe-universal-commerce-protocol-and-why-llms-need-a-new-economic-language.md","The Universal Commerce Protocol and Why LLMs Need a New Economic Language","2026-01-19","Why commerce as we know it does not work for AI and what the Universal Commerce Protocol is trying to fix","\u002Fimages\u002FbannerUCP-blits.png",{},"Why commerce as we know it does not work for AI, what the Universal Commerce Protocol is trying to fix, and why LLMs need a new economic language.","The Universal Commerce Protocol for LLMs","\u002Fblog\u002Fthe-universal-commerce-protocol-and-why-llms-need-a-new-economic-language",{"title":1147,"description":1149},"blog\u002Fthe-universal-commerce-protocol-and-why-llms-need-a-new-economic-language",[],"m0tr5KARy4ydjJJuxAi4PVAYnRO7OhZnTm_iqltnddk",{"id":1160,"title":1161,"author":8,"category":28,"date":1162,"description":1163,"extension":12,"featured":18,"image":1164,"meta":1165,"metaDescription":34,"metaTitle":1166,"navigation":18,"path":1167,"seo":1168,"stem":1169,"tags":1170,"__hash__":1175,"__readTime":88},"blog\u002Fblog\u002Fuse-cases-in-what-the-future-of-banking-ai-actually-looks-like.md","21 Use Cases In: What the Future of Banking AI Actually Looks Like","2026-08-05","21 banking AI use cases, live or in delivery, across the Middle East, Africa and Europe: dialect voice, digital humans, KYC, fraud screening and outbound agents.","\u002Fblog\u002Fblits-banking-ai-hero.png",{},"21 Use Cases: The Future of Banking AI","\u002Fblog\u002Fuse-cases-in-what-the-future-of-banking-ai-actually-looks-like",{"title":1161,"description":1163},"blog\u002Fuse-cases-in-what-the-future-of-banking-ai-actually-looks-like",[749,52,1171,268,1172,1173,1174],"digital humans","super apps","outbound voice","use cases","8xwelyyim84wM89UJ3_AwxnBerKWf2IZG1930y4KjcI",{"id":1177,"title":1178,"author":8,"category":9,"date":1179,"description":1180,"extension":12,"featured":13,"image":1181,"meta":1182,"metaDescription":1183,"metaTitle":1184,"navigation":18,"path":1185,"seo":1186,"stem":1187,"tags":1188,"__hash__":1193,"__readTime":103},"blog\u002Fblog\u002Fusing-cloud-ai-without-sending-personal-data-abroad.md","Using Cloud AI Without Sending Personal Data Abroad","2026-10-16","Strict data rules, no local AI cloud region, no GPUs, and a board that wants agents live this quarter. For a lot of banks and other regulated organizations that used to mean picking two out of three. Our new privacy shield removes the trade off: the best cloud models do the thinking, and the personal data never leaves your building.","\u002Fblog\u002Fblits-privacy-shield-hero.png",{},"Banks and other regulated organizations without a local AI cloud or GPUs can use the best cloud models while personal data stays in their own environment.","Cloud AI Without Sending Personal Data Abroad","\u002Fblog\u002Fusing-cloud-ai-without-sending-personal-data-abroad",{"title":1178,"description":1180},"blog\u002Fusing-cloud-ai-without-sending-personal-data-abroad",[1189,1190,1191,917,1192,52],"privacy","data residency","on premises","regulated industries","6RznfjQ5ouJ8-p4PHDvoYFz7sJc2d3OrPwHh9ZsvWek",{"id":1195,"title":1196,"author":8,"category":28,"date":1197,"description":1198,"extension":12,"featured":13,"image":1199,"meta":1200,"metaDescription":1201,"metaTitle":1202,"navigation":18,"path":1203,"seo":1204,"stem":1205,"tags":1206,"__hash__":1210,"__readTime":39},"blog\u002Fblog\u002Fwe-read-650-ai-deployments-so-you-dont-have-to.md","We Read 1,300+ AI Deployments So You Don't Have To","2026-10-13","We built a public library of 257 AI use cases and 1,377 real deployments, with every number quoted from its source and labelled with who made the claim. The verdict: AI is doing real work, the industry can't prove it, and the sellers are writing the benchmarks.","\u002Fblog\u002Fblits-use-case-library-hero.png",{},"A public library of 257 AI use cases and 1,377 real deployments, every number sourced. AI does real work, but the industry cannot prove it yet.","Lessons From 1,300+ AI Deployments","\u002Fblog\u002Fwe-read-650-ai-deployments-so-you-dont-have-to",{"title":1196,"description":1198},"blog\u002Fwe-read-650-ai-deployments-so-you-dont-have-to",[1000,1207,53,1208,1209],"ai adoption","ai benchmarks","research","1t8kKQ1aRAV876lRNXk4uMDvFc_IGXhaAaVoH9WtrnI",{"id":1212,"title":1213,"author":8,"category":28,"date":937,"description":1214,"extension":12,"featured":13,"image":1215,"meta":1216,"metaDescription":34,"metaTitle":1217,"navigation":18,"path":1218,"seo":1219,"stem":1220,"tags":1221,"__hash__":1224,"__readTime":39},"blog\u002Fblog\u002Fwere-not-saas-not-an-integrator-on-purpose.md","We're Not SaaS. We're Not an Integrator. On Purpose.","Procurement wants us to be a software vendor or an implementation partner. We are neither, and I think the model in between is what gets AI into production.","\u002Fblog\u002Fblits-positioning-hero.png",{},"Not SaaS, Not an Integrator, On Purpose","\u002Fblog\u002Fwere-not-saas-not-an-integrator-on-purpose",{"title":1213,"description":1214},"blog\u002Fwere-not-saas-not-an-integrator-on-purpose",[53,54,1222,1223],"platform","company","NNBRDZDagTR5umxOvb2Os1ZCgfhPE4FIwUzmjgyZSYc",{"id":1226,"title":1227,"author":136,"category":28,"date":1228,"description":1229,"extension":12,"featured":13,"image":1230,"meta":1231,"metaDescription":1232,"metaTitle":1233,"navigation":18,"path":1234,"seo":1235,"stem":1236,"tags":1237,"__hash__":1239,"__readTime":71},"blog\u002Fblog\u002Fwhat-ai-native-means-for-a-bank.md","What \"AI Native\" Actually Means for a Bank","2026-11-13","Every bank now has AI. Very few are AI native. The difference is not the number of pilots but whether the bank's processes run on AI, with people making the decisions. What AI native means in practice, what the leaders already show, and five shifts that separate AI added from AI native.","\u002Fblog\u002Fblits-ai-native-bank-hero.png",{},"AI native is more than a chatbot. What an AI native bank looks like, what DBS, JPMorgan and BNY already do, and five shifts from AI added to AI native.","What AI Native Actually Means for a Bank","\u002Fblog\u002Fwhat-ai-native-means-for-a-bank",{"title":1227,"description":1229},"blog\u002Fwhat-ai-native-means-for-a-bank",[1238,56,54,52,118],"ai native","xm9stxevxfXhA2IiQw26274lTFWljmNkQ1f3_5HznR8",{"id":1241,"title":1242,"author":8,"category":28,"date":1243,"description":1244,"extension":12,"featured":13,"image":1245,"meta":1246,"metaDescription":1247,"metaTitle":1248,"navigation":18,"path":1249,"seo":1250,"stem":1251,"tags":34,"__hash__":1252,"__readTime":151},"blog\u002Fblog\u002Fwhat-is-artificial-intelligence.md","At this point, most people are afraid to ask a simple question; what is artificial intelligence (AI)?","2023-03-28","Someone once told me stupid questions don't exist. Maybe so, but people are still afraid to look stupid by asking about things that are assumed to be general knowledge, and Artificial Intelligence (AI) is one of them. With ChatGPT and Midjourney everyone loves to talk about AI, yet I only know a handful of people who actually know how it works. So here are the basics.","\u002Fblog\u002F1678649727454.jpeg",{},"Everyone talks about ChatGPT and Midjourney, yet few people know how AI actually works, and many are afraid to ask. So here are the basics.","What Is Artificial Intelligence (AI)?","\u002Fblog\u002Fwhat-is-artificial-intelligence",{"title":1242,"description":1244},"blog\u002Fwhat-is-artificial-intelligence","puXtdqK7g8r2GrDbxm4d77lDflWUdtOTBEqLRtvmeIs",[21,37,50,69,82,100,130,246,283,296,314,331,346,705,720,734,747,761,774,792,805,818,831,847,881,897,913,928,943,957,983,1012,1026,1039,1054,1066,1081,1097,1111,1126,1140,1156,1169,1187,1205,1220,1251],1792579736738]