[{"data":1,"prerenderedAt":383},["ShallowReactive",2],{"case-studies-index":3},[4,152,263],{"id":5,"title":6,"body":7,"client":122,"description":123,"extension":124,"industry":125,"meta":126,"metrics":127,"navigation":137,"ogDescription":138,"ogTitle":138,"order":139,"path":140,"relatedService":141,"seo":142,"stack":143,"stem":149,"summary":150,"__hash__":151},"caseStudies\u002Fcase-studies\u002Fai-chatbot-for-customer-support.md","AI Chatbot for Customer Support — Case Study",{"type":8,"value":9,"toc":112},"minimark",[10,14,19,23,29,33,41,61,64,68,101,105,108],[11,12],"case-study-hero",{":breadcrumbs":13},"[{\"label\":\"Home\",\"to\":\"\u002F\"},{\"label\":\"Case Studies\",\"to\":\"\u002Fcase-studies\"},{\"label\":\"AI Support Assistant\"}]",[15,16,18],"h2",{"id":17},"the-challenge","The challenge",[20,21,22],"p",{},"A growing B2B SaaS support team was answering the same product questions over and\nover. Knowledge was spread across a help center, internal policy docs, and the\nheads of a few senior agents. New hires took weeks to ramp, and response times\nclimbed during busy periods.",[24,25,26],"blockquote",{},[20,27,28],{},"The numbers below are representative of this type of engagement. Real figures\nare shared with prospective clients under NDA.",[15,30,32],{"id":31},"what-we-built","What we built",[20,34,35,36,40],{},"We built a ",[37,38,39],"strong",{},"retrieval-augmented knowledge assistant"," that:",[42,43,44,48,55,58],"ul",{},[45,46,47],"li",{},"Ingested the public help center, internal SOPs, and past resolved tickets",[45,49,50,51,54],{},"Answered questions ",[37,52,53],{},"with citations"," back to the source document",[45,56,57],{},"Enforced which content was customer-facing vs. internal-only",[45,59,60],{},"Surfaced \"answer gaps\" — questions the knowledge base couldn't answer well",[20,62,63],{},"The assistant was embedded both in the customer help widget and in the agents'\ninternal console.",[15,65,67],{"id":66},"the-approach","The approach",[69,70,71,77,83,89,95],"ol",{},[45,72,73,76],{},[37,74,75],{},"Content audit"," to find authoritative, up-to-date sources",[45,78,79,82],{},[37,80,81],{},"Ingestion pipeline"," with scheduled re-syncing",[45,84,85,88],{},[37,86,87],{},"Retrieval + re-ranking"," tuned against a labeled evaluation set",[45,90,91,94],{},[37,92,93],{},"Guardrails"," so the assistant defers instead of guessing",[45,96,97,100],{},[37,98,99],{},"Analytics"," on deflection, satisfaction, and content gaps",[15,102,104],{"id":103},"results","Results",[20,106,107],{},"By grounding answers in real content and measuring quality before launch, the\nassistant deflected a meaningful share of repetitive tickets and cut first\nresponse times — while giving agents faster, cited answers.",[15,109,111],{"id":110},"stack","Stack",{"title":113,"searchDepth":114,"depth":114,"links":115},"",3,[116,118,119,120,121],{"id":17,"depth":117,"text":18},2,{"id":31,"depth":117,"text":32},{"id":66,"depth":117,"text":67},{"id":103,"depth":117,"text":104},{"id":110,"depth":117,"text":111},"B2B SaaS company","How we built a RAG knowledge assistant that deflected repetitive support tickets and cut average response time for a SaaS support team.","md","Software \u002F SaaS",{},[128,131,134],{"label":129,"value":130},"Repetitive tickets deflected","~40%",{"label":132,"value":133},"Avg. first response time","−65%",{"label":135,"value":136},"Launch timeline","6 weeks",true,null,1,"\u002Fcase-studies\u002Fai-chatbot-for-customer-support","\u002Fservices\u002Frag-chatbot-development",{"title":6,"description":123},[144,145,146,147,148],"OpenAI","pgvector","Node.js","Nuxt","AWS (Lambda, S3, CloudFront)","case-studies\u002Fai-chatbot-for-customer-support","A RAG knowledge assistant grounded in the company's help center and policies, deflecting repetitive tickets and giving agents instant, cited answers.","zDlco-Jp0kd3zakd8lGPP7uPGjfIzZYHHt4_KYurcrs",{"id":153,"title":154,"body":155,"client":239,"description":240,"extension":124,"industry":241,"meta":242,"metrics":243,"navigation":137,"ogDescription":138,"ogTitle":138,"order":117,"path":253,"relatedService":254,"seo":255,"stack":256,"stem":260,"summary":261,"__hash__":262},"caseStudies\u002Fcase-studies\u002Fcustom-crm-for-service-business.md","Custom CRM for a Service Business — Case Study",{"type":8,"value":156,"toc":232},[157,160,162,165,169,171,177,191,193,225,227,230],[11,158],{":breadcrumbs":159},"[{\"label\":\"Home\",\"to\":\"\u002F\"},{\"label\":\"Case Studies\",\"to\":\"\u002Fcase-studies\"},{\"label\":\"Custom CRM\"}]",[15,161,18],{"id":17},[20,163,164],{},"A multi-location service business ran its operations on a stack of spreadsheets, a\nshared inbox, and manual copy-paste between tools. Leadership had no real-time\nview of the pipeline, and month-end reporting took a senior team member hours of\nerror-prone work.",[24,166,167],{},[20,168,28],{},[15,170,32],{"id":31},[20,172,173,174,40],{},"A ",[37,175,176],{},"custom CRM and reporting portal",[42,178,179,182,185,188],{},[45,180,181],{},"Unified five disconnected data sources into one system",[45,183,184],{},"Managed clients, jobs, and pipeline with role-based access",[45,186,187],{},"Automated month-end reporting that used to be manual",[45,189,190],{},"Gave leadership a live dashboard across all locations",[15,192,67],{"id":66},[69,194,195,201,207,213,219],{},[45,196,197,200],{},[37,198,199],{},"Process mapping"," with the operations team",[45,202,203,206],{},[37,204,205],{},"Data migration"," and cleanup from spreadsheets and legacy tools",[45,208,209,212],{},[37,210,211],{},"Phased rollout"," starting with the highest-pain workflow",[45,214,215,218],{},[37,216,217],{},"Training"," so the team adopted it quickly",[45,220,221,224],{},[37,222,223],{},"Iteration"," based on real usage",[15,226,104],{"id":103},[20,228,229],{},"Reporting that used to take hours became a few minutes, data lived in one place,\nand leadership finally had real-time visibility — without per-seat SaaS costs.",[15,231,111],{"id":110},{"title":113,"searchDepth":114,"depth":114,"links":233},[234,235,236,237,238],{"id":17,"depth":117,"text":18},{"id":31,"depth":117,"text":32},{"id":66,"depth":117,"text":67},{"id":103,"depth":117,"text":104},{"id":110,"depth":117,"text":111},"Multi-location service business","How we replaced spreadsheets and email with a custom CRM and reporting portal for a multi-location service business, cutting manual reporting time dramatically.","Professional services",{},[244,247,250],{"label":245,"value":246},"Manual report generation","hours → minutes",{"label":248,"value":249},"Data sources unified","5 → 1",{"label":251,"value":252},"First release","7 weeks","\u002Fcase-studies\u002Fcustom-crm-for-service-business","\u002Fservices\u002Fcustom-crm-development",{"title":154,"description":240},[257,146,258,259],"Vue \u002F Nuxt","PostgreSQL","AWS (ECS, RDS, S3)","case-studies\u002Fcustom-crm-for-service-business","A custom CRM and reporting portal that replaced spreadsheets and shared inboxes, automating reporting and giving leadership real-time visibility.","966Y4dIuSjnbN0Ge6AOK65H2XGu3IGmqqXu-QMXxBMM",{"id":264,"title":265,"body":266,"client":362,"description":363,"extension":124,"industry":364,"meta":365,"metrics":366,"navigation":137,"ogDescription":138,"ogTitle":138,"order":114,"path":375,"relatedService":376,"seo":377,"stack":378,"stem":380,"summary":381,"__hash__":382},"caseStudies\u002Fcase-studies\u002Fai-product-assistant-ecommerce.md","AI Product Assistant for an Online Store — Case Study",{"type":8,"value":267,"toc":355},[268,271,273,276,280,282,293,310,313,315,345,347,350,352],[11,269],{":breadcrumbs":270},"[{\"label\":\"Home\",\"to\":\"\u002F\"},{\"label\":\"Case Studies\",\"to\":\"\u002Fcase-studies\"},{\"label\":\"AI Product Assistant\"}]",[15,272,18],{"id":17},[20,274,275],{},"An online store with a deep, specialized catalog was losing shoppers who couldn't\ntell which product was right for them. The differences between items were nuanced,\nbuyers asked the same pre-sale questions over and over, and the team couldn't\nanswer fast enough to keep every visitor from bouncing.",[24,277,278],{},[20,279,28],{},[15,281,32],{"id":31},[20,283,284,285,288,289,292],{},"We embedded an ",[37,286,287],{},"AI shopping assistant"," in the storefront that can consult\ncustomers on ",[37,290,291],{},"any product in the catalog",". It:",[42,294,295,298,301,304,307],{},[45,296,297],{},"Explains what each product is and how products differ",[45,299,300],{},"Recommends the right item for a stated goal or use case",[45,302,303],{},"Compares options side by side in plain language",[45,305,306],{},"Cites the exact product pages its answers come from",[45,308,309],{},"Stays inside catalog data — so it never invents specs or makes off-label claims",[20,311,312],{},"The assistant is available on the homepage and on every product page, pre-loaded\nwith the context of the item the shopper is viewing.",[15,314,67],{"id":66},[69,316,317,323,329,334,340],{},[45,318,319,322],{},[37,320,321],{},"Catalog ingestion"," — products, categories, and descriptions indexed for retrieval",[45,324,325,328],{},[37,326,327],{},"Retrieval + grounding"," so answers stay tied to real products",[45,330,331,333],{},[37,332,93],{}," that keep claims accurate and compliant",[45,335,336,339],{},[37,337,338],{},"Buy-flow handoff"," — the assistant links straight to add-to-cart",[45,341,342,344],{},[37,343,99],{}," on questions asked, products recommended, and gaps",[15,346,104],{"id":103},[20,348,349],{},"Shoppers could self-serve product guidance instead of waiting on the team, pre-sale\nquestions dropped sharply, and the assistant nudged buyers toward the right product\nwith sourced, on-brand answers across the entire catalog.",[15,351,111],{"id":110},[20,353,354],{},"A retrieval-augmented assistant over the live catalog, deployed on cloud\ninfrastructure and embedded directly in the headless storefront.",{"title":113,"searchDepth":114,"depth":114,"links":356},[357,358,359,360,361],{"id":17,"depth":117,"text":18},{"id":31,"depth":117,"text":32},{"id":66,"depth":117,"text":67},{"id":103,"depth":117,"text":104},{"id":110,"depth":117,"text":111},"Specialty ecommerce store","How we added an AI shopping assistant to an ecommerce catalog that consults customers on any product, grounded in the store's real data — boosting product discovery and reducing pre-sale questions.","Ecommerce \u002F Retail",{},[367,370,373],{"label":368,"value":369},"Catalog covered by assistant","100%",{"label":371,"value":372},"Pre-sale product questions","−50%",{"label":135,"value":374},"5 weeks","\u002Fcase-studies\u002Fai-product-assistant-ecommerce","\u002Fservices\u002Fecommerce-development",{"title":265,"description":363},[144,145,147,379,148],"Headless commerce","case-studies\u002Fai-product-assistant-ecommerce","An AI shopping assistant embedded in an online store that consults customers on any product — explaining items, comparing options, and recommending the right fit, grounded in the store's own catalog.","Rs9-gUE7S237Jq0YrKAwDjFyj-0Bi6Mi1wNJ9hIrSCk",1782707468355]