[{"data":1,"prerenderedAt":226},["ShallowReactive",2],{"work-list":3,"insights-list":132},[4,87],{"id":5,"title":6,"body":7,"client":72,"description":73,"extension":74,"featured":75,"meta":76,"navigation":75,"path":77,"seo":78,"services":79,"stem":84,"year":85,"__hash__":86},"work\u002Fwork\u002Fcommerce-operations-platform.md","From fragmented fulfilment workflows to an integrated operational platform",{"type":8,"value":9,"toc":63},"minimark",[10,15,19,23,26,30,33,37,56,60],[11,12,14],"h2",{"id":13},"context","Context",[16,17,18],"p",{},"A growing commerce operation had accumulated business-critical fulfilment logic across multiple systems, integrations and manual procedures.",[11,20,22],{"id":21},"problem","Problem",[16,24,25],{},"The storefront was only one part of the operation. Order routing, inventory state, labels, carrier rules and warehouse processes had become tightly coupled, making seemingly small changes operationally risky.",[11,27,29],{"id":28},"nekomatas-role","Nekomata's role",[16,31,32],{},"Nekomata provided architecture, implementation and long-term technical ownership across the operational stack.",[11,34,36],{"id":35},"approach","Approach",[38,39,40,44,47,50,53],"ul",{},[41,42,43],"li",{},"Identify the operational source of truth for each critical state.",[41,45,46],{},"Make system boundaries and failure modes explicit.",[41,48,49],{},"Reduce manual intervention where automation is reliable.",[41,51,52],{},"Preserve deliberate human control where exceptions carry business risk.",[41,54,55],{},"Treat warehouse output and shipping artefacts as first-class system responsibilities.",[11,57,59],{"id":58},"outcome","Outcome",[16,61,62],{},"Replace this section with the outcomes you are permitted to disclose: reduced manual handling, faster fulfilment, lower failure rates, increased throughput, operational visibility, or longevity of the platform.",{"title":64,"searchDepth":65,"depth":65,"links":66},"",2,[67,68,69,70,71],{"id":13,"depth":65,"text":14},{"id":21,"depth":65,"text":22},{"id":28,"depth":65,"text":29},{"id":35,"depth":65,"text":36},{"id":58,"depth":65,"text":59},"Anonymised commerce client","Architecture and long-term technical ownership across commerce, orders, inventory, fulfilment and warehouse tooling.","md",true,{},"\u002Fwork\u002Fcommerce-operations-platform",{"title":6,"description":73},[80,81,82,83],"Commerce architecture","OMS \u002F WMS integration","Fulfilment","ZPL & carrier tooling","work\u002Fcommerce-operations-platform",2026,"hoOb6hJfY5D45cDtjjLJZ2Iq_I7D13GjzGnzsA5SQOU",{"id":88,"title":89,"body":90,"client":121,"description":122,"extension":74,"featured":75,"meta":123,"navigation":75,"path":124,"seo":125,"services":126,"stem":130,"year":85,"__hash__":131},"work\u002Fwork\u002Fplatform-modernisation.md","Modernising a business-critical inherited platform without stopping operations",{"type":8,"value":91,"toc":115},[92,94,97,99,102,104,107,110,112],[11,93,14],{"id":13},[16,95,96],{},"The platform had become essential to daily operations while documentation and architectural clarity had not kept pace with its importance.",[11,98,22],{"id":21},[16,100,101],{},"A complete rewrite would have introduced unacceptable delivery and operational risk. Continuing without intervention made every future change slower and harder to reason about.",[11,103,36],{"id":35},[16,105,106],{},"The engagement begins by reconstructing the actual system: data flows, ownership, dependencies, failure behaviour and operational workarounds.",[16,108,109],{},"Modernisation is then staged around risk rather than fashion.",[11,111,59],{"id":58},[16,113,114],{},"Use this starter study to document a real rescue or inherited-system engagement once you choose the example you want to publish.",{"title":64,"searchDepth":65,"depth":65,"links":116},[117,118,119,120],{"id":13,"depth":65,"text":14},{"id":21,"depth":65,"text":22},{"id":35,"depth":65,"text":36},{"id":58,"depth":65,"text":59},"Anonymised platform client","Technical discovery, risk mapping and staged modernisation for a platform that could not simply be rebuilt from scratch.",{},"\u002Fwork\u002Fplatform-modernisation",{"title":89,"description":122},[127,128,129],"Technical discovery","Legacy modernisation","Architecture","work\u002Fplatform-modernisation","hbSr7elTrKKHj6tHpezSn_PfDy5Qc0A5HaKUC63gAaE",[133,182],{"id":134,"title":135,"body":136,"category":174,"date":175,"description":176,"extension":74,"meta":177,"navigation":75,"path":178,"seo":179,"stem":180,"__hash__":181},"insights\u002Finsights\u002Fstorefront-boundary.md","Why commerce architecture stops at the wrong boundary",{"type":8,"value":137,"toc":171},[138,141,144,161,165,168],[16,139,140],{},"A storefront is easy to draw on an architecture diagram because its boundary is visually obvious. Operational commerce is not.",[16,142,143],{},"The difficult questions often begin after a customer has already paid:",[38,145,146,149,152,155,158],{},[41,147,148],{},"Which system owns the order now?",[41,150,151],{},"Which inventory state is authoritative?",[41,153,154],{},"What happens when fulfilment can only be partially completed?",[41,156,157],{},"Who can safely retry a failed integration?",[41,159,160],{},"Which system is allowed to print the shipping artefact?",[11,162,164],{"id":163},"architecture-should-follow-responsibility","Architecture should follow responsibility",[16,166,167],{},"A useful architecture map does not merely show APIs. It shows who owns decisions and state.",[16,169,170],{},"That distinction is what turns a collection of integrations into an operable system.",{"title":64,"searchDepth":65,"depth":65,"links":172},[173],{"id":163,"depth":65,"text":164},"Commerce Architecture","2026-08-14","The storefront is visible, but the expensive failures often happen after checkout.",{},"\u002Finsights\u002Fstorefront-boundary",{"title":135,"description":176},"insights\u002Fstorefront-boundary","minPigi9A4w8Ufsz8fIPwCvYkLA110gkUEp6SAcyLOo",{"id":183,"title":184,"body":185,"category":218,"date":219,"description":220,"extension":74,"meta":221,"navigation":75,"path":222,"seo":223,"stem":224,"__hash__":225},"insights\u002Finsights\u002Fai-operations.md","AI becomes useful when it enters an actual operating model",{"type":8,"value":186,"toc":216},[187,190,193,213],[16,188,189],{},"The easiest AI prototype is often the least interesting production system.",[16,191,192],{},"A useful operational implementation needs answers to ordinary architecture questions:",[38,194,195,198,201,204,207,210],{},[41,196,197],{},"What data may the model see?",[41,199,200],{},"Which actions may it perform?",[41,202,203],{},"Which actions require approval?",[41,205,206],{},"What happens when a tool fails?",[41,208,209],{},"How is a decision reconstructed later?",[41,211,212],{},"Where does deterministic software remain the better choice?",[16,214,215],{},"The model is only one component. The operating model around it is what makes the capability trustworthy.",{"title":64,"searchDepth":65,"depth":65,"links":217},[],"Applied AI","2026-08-10","Connecting intelligent tooling to business systems requires more architecture, not less.",{},"\u002Finsights\u002Fai-operations",{"title":184,"description":220},"insights\u002Fai-operations","hOkgiJMshOn-1zHlv8wmElJB07j4E9rr0U6EzXPWCsI",1787213493714]