{"id":122615,"date":"2026-11-03T10:45:38","date_gmt":"2026-11-03T09:45:38","guid":{"rendered":"https:\/\/aixia.se\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/"},"modified":"2026-11-03T10:45:48","modified_gmt":"2026-11-03T09:45:48","slug":"from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026","status":"publish","type":"post","link":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/","title":{"rendered":"From Model to Production: LLMOps Will Be the Most Important MLOps Shift of 2026"},"content":{"rendered":"<div class=\"aixia-post\">\n<style>\n.aixia-post { font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif; color: #212529; line-height: 1.65; max-width: 820px; margin: 0 auto; }\n.aixia-post p.lead { font-size: 1.18em; color: #003366; line-height: 1.6; }\n.aixia-post h2 { font-size: 21px; color: #003366; margin: 34px 0 10px; padding-bottom: 6px; border-bottom: 2px solid #0066CC; }\n.aixia-post h3 { font-size: 17px; color: #003366; margin: 24px 0 8px; }\n.aixia-post table { width: 100%; border-collapse: collapse; margin: 18px 0; font-size: 15px; }\n.aixia-post th { background: #eaf3fb; color: #003366; text-align: left; padding: 10px 12px; border: 1px solid #d6e4f0; }\n.aixia-post td { padding: 10px 12px; border: 1px solid #e4edf5; vertical-align: top; }\n.aixia-post tr:nth-child(even) td { background: #f8fbfd; }\n.aixia-post.cta { background: #f0f6fc; border: 1px solid #cfe0ef; border-left: 4px solid #0066CC; border-radius: 6px; padding: 20px 22px; margin: 30px 0 6px; }\n.aixia-post.cta strong { color: #003366; }\n.aixia-post.kallor { font-size: 13px; color: #6c757d; margin-top: 26px; border-top: 1px solid #e4edf5; padding-top: 12px; }\n.aixia-post.kallor li { margin-bottom: 4px; }\n<\/style>\n<p class=\"lead\">En AI-assistent som fungerar i en demo \u00e4r en bra b\u00f6rjan. N\u00e4r den ska anv\u00e4ndas av hela organisationen beh\u00f6ver n\u00e5gon ocks\u00e5 ansvara f\u00f6r svarskvalitet, beh\u00f6righeter, uppdateringar och kostnader. V\u00e5r bed\u00f6mning \u00e4r att just den f\u00f6rflyttningen g\u00f6r LLMOps till ett av 2026 \u00e5rs viktigaste skiften inom MLOps.<\/p>\n<p>En intern kunskapsassistent svarar snabbt och korrekt i piloten. Sedan ansluts fler dokument, fler avdelningar och fler verksamhetssystem. Anv\u00e4ndarna st\u00e4ller fr\u00e5gor som utvecklarna inte f\u00f6ruts\u00e5g. Svarstiderna varierar och kostnaden v\u00e4xer. Ibland blir svaret fel, trots att alla servrar visar gr\u00f6nt.<\/p>\n<p>Det \u00e4r h\u00e4r LLMOps blir avg\u00f6rande: arbetss\u00e4tten och verktygen f\u00f6r att utveckla, drifts\u00e4tta, f\u00f6lja upp och f\u00f6rvalta tj\u00e4nster som bygger p\u00e5 stora spr\u00e5kmodeller.<\/p>\n<p>F\u00f6r f\u00f6retag som vill anv\u00e4nda generativ AI i ordinarie verksamhet beh\u00f6ver detta finnas med redan f\u00f6re produktionsstart. Annars riskerar varje lyckad pilot att bli en tj\u00e4nst som ingen riktigt \u00e4ger.<\/p>\n<h2>Varf\u00f6r fr\u00e5gan blir viktigare under 2026<\/h2>\n<p>Gartner f\u00f6rutsp\u00e5dde i mars 2024 att AI och verktyg som anv\u00e4nder stora spr\u00e5kmodeller skulle st\u00e5 f\u00f6r \u00f6ver 30 procent av \u00f6kningen i efterfr\u00e5gan p\u00e5 API:er fram till 2026. Det \u00e4r en prognos om efterfr\u00e5getillv\u00e4xt, inte en uppm\u00e4tt andel av alla API-anrop under \u00e5ret.<\/p>\n<p>I en prognos fr\u00e5n juli 2026 betonar Gartner dessutom kundernas \u00f6kande fokus p\u00e5 kostnadskontroll, anv\u00e4ndning och m\u00e4tbara resultat inom AI-plattformar och modeller.<\/p>\n<p>F\u00f6r IT-organisationen \u00e4r f\u00f6ljden konkret. N\u00e4r AI blir en del av arbetsfl\u00f6den m\u00e5ste den hanteras som andra verksamhetstj\u00e4nster: med ansvariga \u00e4gare, kontrollerade f\u00f6r\u00e4ndringar och beredskap n\u00e4r n\u00e5got g\u00e5r fel. Samtidigt tillkommer en fr\u00e5ga som traditionell drift\u00f6vervakning inte besvarar: levererar tj\u00e4nsten fortfarande tillr\u00e4ckligt bra svar?<\/p>\n<h2>Vad skiljer LLMOps fr\u00e5n MLOps?<\/h2>\n<p>LLMOps bygger vidare p\u00e5 MLOps. Versionshantering, tester, drifts\u00e4ttning och \u00f6vervakning beh\u00f6vs fortfarande. Skillnaden \u00e4r hur stor del av beteendet som kan ligga utanf\u00f6r sj\u00e4lva modellen.<\/p>\n<p>En LLM-tj\u00e4nst kan best\u00e5 av en spr\u00e5kmodell, systeminstruktioner, dokumenth\u00e4mtning, \u00e5tkomstregler och verktyg som anropar andra system. RAG, retrieval-augmented generation, inneb\u00e4r att tj\u00e4nsten h\u00e4mtar relevant information som underlag f\u00f6r modellens svar. \u00c4ven den h\u00e4mtningen m\u00e5ste testas och f\u00f6rvaltas.<\/p>\n<table>\n<thead>\n<tr>\n<th>Omr\u00e5de<\/th>\n<th>Vad LLMOps beh\u00f6ver omfatta<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><strong>Versioner<\/strong><\/td>\n<td>Modell, promptar, applikationskod, konfiguration, dokumentindex och verktygsdefinitioner<\/td>\n<\/tr>\n<tr>\n<td><strong>Kvalitet<\/strong><\/td>\n<td>Om svaret \u00e4r relevant, har st\u00f6d i r\u00e4tt k\u00e4llor och f\u00f6ljer verksamhetens krav<\/td>\n<\/tr>\n<tr>\n<td><strong>S\u00e4kerhet<\/strong><\/td>\n<td>Vilka data tj\u00e4nsten f\u00e5r l\u00e4sa och vilka handlingar den f\u00e5r utf\u00f6ra f\u00f6r varje anv\u00e4ndare<\/td>\n<\/tr>\n<tr>\n<td><strong>F\u00f6r\u00e4ndringar<\/strong><\/td>\n<td>Tester \u00e4ven n\u00e4r promptar, dokumenth\u00e4mtning eller modellleverant\u00f6r \u00e4ndras<\/td>\n<\/tr>\n<tr>\n<td><strong>Ekonomi<\/strong><\/td>\n<td>Kostnad per godk\u00e4nt resultat, inklusive \u00e5terf\u00f6rs\u00f6k och m\u00e4nsklig efterbearbetning<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>Mycket av detta \u00e4r ocks\u00e5 relevant f\u00f6r andra ML-system. LLMOps g\u00f6r omfattningen s\u00e4rskilt tydlig: den enhet som ska f\u00f6rvaltas \u00e4r hela AI-tj\u00e4nsten. Om ni redan har en MLOps-praktik p\u00e5 plats \u00e4r den ett bra fundament; de konkreta fallgroparna i steget fr\u00e5n pilot till produktion har vi samlat i <a href=\"https:\/\/aixia.se\/en\/from-pilot-to-production-the-mlops-pitfalls-when-scaling-ai\/\">Fr\u00e5n pilot till produktion: MLOps-f\u00e4llorna n\u00e4r AI ska skalas<\/a>.<\/p>\n<h2>\u00d6vervakningen m\u00e5ste visa mer \u00e4n tillg\u00e4nglighet<\/h2>\n<p>Ett API kan svara utan felkod och \u00e4nd\u00e5 leverera ett oanv\u00e4ndbart svar. D\u00e4rf\u00f6r beh\u00f6ver \u00f6vervakningen t\u00e4cka b\u00e5de teknisk funktion och verksamhetskvalitet.<\/p>\n<p>P\u00e5 den tekniska sidan beh\u00f6vs exempelvis k\u00f6tid, felandel, kapacitetsutnyttjande och svarstid. F\u00f6r str\u00f6mmade LLM-svar \u00e4r tiden till f\u00f6rsta token och tiden f\u00f6r hela svaret olika m\u00e5tt. Svarstiden f\u00f6r de l\u00e5ngsammaste f\u00f6rfr\u00e5gningarna, exempelvis p95, kan vara mer anv\u00e4ndbar \u00e4n ett genomsnitt.<\/p>\n<p>P\u00e5 kvalitetssidan beh\u00f6ver organisationen ett representativt testunderlag och \u00e5terkommande granskning. Besvaras fr\u00e5gan? \u00c4r k\u00e4llorna relevanta och aktuella? St\u00f6der de faktiskt slutsatsen? Kan tj\u00e4nsten l\u00e4mna \u00f6ver till en m\u00e4nniska n\u00e4r underlaget inte r\u00e4cker?<\/p>\n<p>Verktyg som MLflow har st\u00f6d f\u00f6r sp\u00e5rning och utv\u00e4rdering av produktionsfl\u00f6den, inklusive automatisk bed\u00f6mning av ett urval svar. Det kan hj\u00e4lpa team att uppt\u00e4cka avvikelser och f\u00f6lja utvecklingen \u00f6ver tid.<\/p>\n<p>Men automatisk utv\u00e4rdering \u00e4r en signal, inte ett facit. En annan spr\u00e5kmodell som bed\u00f6mer svaret kan ocks\u00e5 g\u00f6ra fel. Kombinera d\u00e4rf\u00f6r automatiska kontroller med verksamhetens referensfall och m\u00e4nsklig granskning. Kritiska fel f\u00e5r inte d\u00f6ljas av ett bra genomsnittsv\u00e4rde.<\/p>\n<h3>Model drift: tj\u00e4nsten kan bli s\u00e4mre utan att modellen \u00e4ndras<\/h3>\n<p>Begreppet model drift anv\u00e4nds ofta som ett samlingsnamn f\u00f6r f\u00f6rs\u00e4mrad modellprestanda. I LLM-tj\u00e4nster beh\u00f6ver man skilja mellan orsakerna.<\/p>\n<p>En modell med frysta vikter l\u00e4r sig inte automatiskt av anv\u00e4ndarnas fr\u00e5gor. \u00c4nd\u00e5 kan tj\u00e4nstens kvalitet f\u00f6r\u00e4ndras n\u00e4r andra delar av systemet eller verksamheten \u00e4ndras.<\/p>\n<table>\n<thead>\n<tr>\n<th>F\u00f6r\u00e4ndring<\/th>\n<th>Exempel<\/th>\n<th>M\u00f6jlig \u00e5tg\u00e4rd efter analys<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Anv\u00e4ndningen f\u00f6r\u00e4ndras<\/td>\n<td>En assistent f\u00e5r fr\u00e5gor fr\u00e5n en ny yrkesgrupp<\/td>\n<td>Ut\u00f6ka testfallen och bed\u00f6m om anv\u00e4ndningsomr\u00e5det ska breddas<\/td>\n<\/tr>\n<tr>\n<td>Kunskapsunderlaget f\u00f6r\u00e4ndras<\/td>\n<td>En gammal rutin ligger kvar i dokumentindexet<\/td>\n<td>Uppdatera indexet och pr\u00f6va k\u00e4llornas prioritering<\/td>\n<\/tr>\n<tr>\n<td>Instruktionerna \u00e4ndras<\/td>\n<td>En ny prompt f\u00f6rs\u00e4mrar k\u00e4llh\u00e4nvisningarna<\/td>\n<td>J\u00e4mf\u00f6r versioner och \u00e5terst\u00e4ll vid behov<\/td>\n<\/tr>\n<tr>\n<td>Modell eller runtime byts<\/td>\n<td>En uppgradering p\u00e5verkar kvalitet eller svarstid<\/td>\n<td>K\u00f6r regressionstester och kontrollera konfigurationen<\/td>\n<\/tr>\n<tr>\n<td>Verksamhetens krav \u00e4ndras<\/td>\n<td>Nya regler f\u00f6r vilken information som f\u00e5r l\u00e4mnas ut<\/td>\n<td>Uppdatera \u00e5tkomstkontroller, policyer och testunderlag<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>En avvikelse ska d\u00e4rf\u00f6r leda till fels\u00f6kning innan n\u00e5gon best\u00e4ller mer tr\u00e4ning. Problemet kan ligga i dokumenth\u00e4mtning, beh\u00f6righeter eller en \u00e4ndrad integration. Omskolning eller finjustering \u00e4r en m\u00f6jlig \u00e5tg\u00e4rd f\u00f6rst n\u00e4r analysen visar att det \u00e4r r\u00e4tt v\u00e4g.<\/p>\n<h2>Policy-as-code g\u00f6r styrningen m\u00f6jlig att kontrollera<\/h2>\n<p>AI-policyer f\u00e5r begr\u00e4nsad effekt om de bara ligger i ett dokument. Delar av styrningen beh\u00f6ver bli regler som kan versionshanteras, testas och till\u00e4mpas automatiskt.<\/p>\n<p>Det \u00e4r inneb\u00f6rden av policy-as-code. Verktyg som Open Policy Agent kan anv\u00e4ndas f\u00f6r policybeslut i bland annat CI\/CD, API-gateways och applikationer.<\/p>\n<p>F\u00f6r en LLM-tj\u00e4nst kan det inneb\u00e4ra att en release stoppas om den anv\u00e4nder en modell som inte \u00e4r godk\u00e4nd f\u00f6r dataklassen, saknar obligatoriska tester eller till\u00e5ter trafik till otill\u00e5tna destinationer. Det kan ocks\u00e5 inneb\u00e4ra att ett verktygsanrop nekas eftersom anv\u00e4ndaren saknar r\u00e4tt beh\u00f6righet.<\/p>\n<p>Kontrollerna beh\u00f6vs p\u00e5 flera st\u00e4llen. CI\/CD kontrollerar vad som f\u00e5r sl\u00e4ppas. Tj\u00e4nsten och de anslutna systemen m\u00e5ste ocks\u00e5 kontrollera vad som f\u00e5r h\u00e4nda under k\u00f6rning. En godk\u00e4nd release ger inte alla anv\u00e4ndare r\u00e4tt till alla data.<\/p>\n<p>En instruktion till modellen att &#8220;inte l\u00e4mna ut k\u00e4nslig information&#8221; ers\u00e4tter inte \u00e5tkomstkontroll. Beh\u00f6righeter m\u00e5ste kontrolleras innan information h\u00e4mtas och n\u00e4r en handling ska utf\u00f6ras. OWASP beskriver hur prompt injection och alltf\u00f6r omfattande verktygsbeh\u00f6righeter kan leda till informationsl\u00e4ckage och otill\u00e5tna handlingar.<\/p>\n<p>F\u00f6r AI-agenter blir detta s\u00e4rskilt konkret. En assistent som f\u00f6resl\u00e5r en \u00e4ndring och en agent som genomf\u00f6r den har olika riskniv\u00e5er. Begr\u00e4nsa verktyg, r\u00e4ttigheter och antal steg. L\u00e4gg m\u00e4nskligt godk\u00e4nnande d\u00e4r en handling kan f\u00e5 st\u00f6rre konsekvenser.<\/p>\n<h3>En prompt\u00e4ndring beh\u00f6ver ocks\u00e5 en releaseprocess<\/h3>\n<p>Det ska g\u00e5 att se vad som \u00e4ndrades n\u00e4r kvaliteten f\u00f6ll. D\u00e4rf\u00f6r beh\u00f6ver promptar och konfiguration behandlas med samma disciplin som applikationskod.<\/p>\n<p>Ett praktiskt fl\u00f6de \u00e4r att versionshantera \u00e4ndringen, k\u00f6ra kvalitets- och s\u00e4kerhetstester, kontrollera policyerna och sedan sl\u00e4ppa till en begr\u00e4nsad grupp. Bredare utrullning sker f\u00f6rst n\u00e4r resultat och belastning h\u00e5ller.<\/p>\n<p>Det ska ocks\u00e5 finnas en plan f\u00f6r \u00e5terst\u00e4llning. Att byta tillbaka modellen r\u00e4cker inte om dokumentindex, verktyg eller beroenden samtidigt har \u00e4ndrats. Dokumentera vilka delar som h\u00f6r till releasen och vad som faktiskt g\u00e5r att \u00e5terst\u00e4lla.<\/p>\n<p>Testerna b\u00f6r \u00e4ven inneh\u00e5lla misslyckanden: ett anslutet system svarar inte, en fr\u00e5ga saknar k\u00e4llunderlag eller en anv\u00e4ndare f\u00f6rs\u00f6ker komma \u00e5t n\u00e5gon annans information. Best\u00e4m hur tj\u00e4nsten ska bete sig i dessa fall f\u00f6re lansering.<\/p>\n<h2>Loggning, kostnad och ansvar i en managed tj\u00e4nst<\/h2>\n<p>F\u00f6r att fels\u00f6ka beh\u00f6ver man kunna f\u00f6lja ett anrop genom dokumenth\u00e4mtning, modellk\u00f6rning och eventuella verktyg. Men promptar och svar kan inneh\u00e5lla personuppgifter, kundinformation och interna dokument.<\/p>\n<p>OpenTelemetrys GenAI-konventioner lyfter just den risken och rekommenderar att k\u00e4nsligt inneh\u00e5ll inte samlas in som standard.<\/p>\n<p>En rimlig utg\u00e5ngspunkt \u00e4r att samla tekniska m\u00e4tv\u00e4rden och versionsmetadata, och d\u00e4refter besluta n\u00e4r inneh\u00e5ll verkligen beh\u00f6ver sparas. Anv\u00e4nd maskning, \u00e5tkomstbegr\u00e4nsning, urval och best\u00e4mda lagringstider. Kontrollera ocks\u00e5 vart data skickas om ett externt verktyg anv\u00e4nds f\u00f6r att bed\u00f6ma svarskvalitet.<\/p>\n<p>P\u00e5 kostnadssidan beh\u00f6vs en koppling till anv\u00e4ndare, projekt och verksamhetsnytta. Fler API-anrop kan bero p\u00e5 h\u00f6gre anv\u00e4ndning, men ocks\u00e5 p\u00e5 \u00e5terf\u00f6rs\u00f6k, l\u00e5nga agentkedjor eller misslyckade svar.<\/p>\n<p>F\u00f6lj d\u00e4rf\u00f6r kostnad per godk\u00e4nt \u00e4rende eller slutf\u00f6rd uppgift, inte bara kostnad per token. I egen infrastruktur beh\u00f6ver \u00e4ven kapacitet, drift och nyttjandegrad ing\u00e5. S\u00e4tt budgetar och gr\u00e4nser f\u00f6r token, samtidighet och agentsteg s\u00e5 att kostnadsproblem g\u00e5r att begr\u00e4nsa innan m\u00e5nadsrapporten kommer.<\/p>\n<p>Alla f\u00f6retag beh\u00f6ver inte bygga ett eget team som hanterar varje del av AI-stacken. Men en extern driftpartner beh\u00f6ver ett tydligt uppdrag. Att h\u00e5lla infrastrukturen ig\u00e5ng \u00e4r en annan uppgift \u00e4n att avg\u00f6ra om ett AI-svar \u00e4r aff\u00e4rsm\u00e4ssigt korrekt.<\/p>\n<table>\n<thead>\n<tr>\n<th>Omr\u00e5de<\/th>\n<th>Driftpartnerns uppgift kan omfatta<\/th>\n<th>Kundens ansvar beh\u00f6ver omfatta<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><strong>Plattform<\/strong><\/td>\n<td>Kapacitet, patchning, \u00f6vervakning och \u00e5terst\u00e4llning<\/td>\n<td>Prioriteringar och tillg\u00e4nglighetsbehov<\/td>\n<\/tr>\n<tr>\n<td><strong>Modellreleaser<\/strong><\/td>\n<td>Testk\u00f6rning, drifts\u00e4ttning och teknisk \u00e5terst\u00e4llning<\/td>\n<td>Godk\u00e4nda anv\u00e4ndningsomr\u00e5den och kvalitetsniv\u00e5er<\/td>\n<\/tr>\n<tr>\n<td><strong>Kvalitetsuppf\u00f6ljning<\/strong><\/td>\n<td>M\u00e4tning, avvikelserapporter och fels\u00f6kning<\/td>\n<td>Referenssvar, sakkunskap och verksamhetsbeslut<\/td>\n<\/tr>\n<tr>\n<td><strong>Styrning<\/strong><\/td>\n<td>Implementering och uppf\u00f6ljning av beslutade kontroller<\/td>\n<td>Informationsklassning, policybeslut och riskacceptans<\/td>\n<\/tr>\n<tr>\n<td><strong>Ekonomi<\/strong><\/td>\n<td>Kostnadsrapportering och kapacitetsoptimering<\/td>\n<td>Budget och bed\u00f6mning av verksamhetsnytta<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>Tillg\u00e4nglighet och svarstid kan f\u00e5 tydliga serviceniv\u00e5er. Kvalitet beh\u00f6ver ocks\u00e5 m\u00e4tbara m\u00e5l, men inom ett definierat anv\u00e4ndningsomr\u00e5de och mot ett \u00f6verenskommet testunderlag. Ett driftavtal kan inte lova att varje genererat svar alltid \u00e4r sant.<\/p>\n<h2>Aixias roll: fr\u00e5n infrastruktur till f\u00f6rvaltning<\/h2>\n<p>Aixia AB f\u00f6renar <a href=\"https:\/\/aixia.se\/en\/our-offer\/ai-solutions-2\/ai-infrastructure-2\/\">AI-infrastruktur<\/a> med plattform och driftkompetens. <a href=\"https:\/\/aixia.se\/en\/our-offer\/ai-solutions-2\/aiqu\/\">AiQu<\/a> ger st\u00f6d f\u00f6r att k\u00f6ra och hantera AI-arbetslaster, dela resurser mellan team och f\u00f6lja resursanv\u00e4ndning. Plattformen har alternativ f\u00f6r lokal och hybrid drifts\u00e4ttning, och beskriver bland annat kostnadsuppf\u00f6ljning per anv\u00e4ndare och projekt.<\/p>\n<p>Det ger en grund f\u00f6r produktionssatt AI. F\u00f6r att m\u00f6ta kundens krav beh\u00f6ver den kompletteras med r\u00e4tt modellutv\u00e4rdering, s\u00e4kerhetskontroller och processer f\u00f6r f\u00f6r\u00e4ndringar och incidenter. H\u00e4r \u00e4r <a href=\"https:\/\/aixia.se\/en\/our-offer\/it-drift-2\/\">IT Drift som tj\u00e4nst (Managed Services)<\/a> navet: samma driftdisciplin som f\u00f6r andra aff\u00e4rskritiska system, men med svar och kvalitet som m\u00e4tbara storheter.<\/p>\n<p>V\u00e5r utg\u00e5ngspunkt \u00e4r att b\u00f6rja med en konkret tj\u00e4nst: vilka anv\u00e4nder den, vilka data f\u00e5r den hantera och vad m\u00e5ste den prestera? D\u00e4refter dimensionerar vi infrastrukturen och definierar hur drift, uppf\u00f6ljning och f\u00f6rb\u00e4ttring ska fungera tillsammans med kundens organisation.<\/p>\n<div class=\"cta\">\n<p><strong>Har ni en LLM-pilot som ska bli en f\u00f6rvaltad verksamhetstj\u00e4nst?<\/strong> Kontakta Aixia AB f\u00f6r att g\u00e5 igenom produktionskrav, ansvarsf\u00f6rdelning och hur en managed MLOps- och LLMOps-l\u00f6sning kan utformas f\u00f6r er milj\u00f6.<\/p>\n<\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>Here\u2019s how to take language models from pilot to production with LLMOps: monitoring responses\u2014not just servers\u2014model operations, policy-as-code, and distributed responsibility in a managed service.<\/p>\n","protected":false},"author":15,"featured_media":122611,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"om_disable_all_campaigns":false,"inline_featured_image":false,"_monsterinsights_skip_tracking":false,"footnotes":""},"categories":[60],"tags":[],"class_list":["post-122615","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog"],"acf":[],"aioseo_notices":[],"aioseo_head":"\n\t\t<!-- All in One SEO 5.0.4 - aioseo.com -->\n\t<meta name=\"description\" content=\"Here&#039;s how to take language models from pilot to production with LLMOps. Learn about monitoring, model operations, policy-as-code, and managed MLOps with Aixia.\" \/>\n\t<meta name=\"robots\" content=\"max-image-preview:large\" \/>\n\t<meta name=\"author\" content=\"Aixia\"\/>\n\t<link rel=\"canonical\" href=\"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/\" \/>\n\t\t<meta name=\"generator\" content=\"All in One SEO (AIOSEO) 5.0.4\" \/>\n\t\t<meta property=\"og:locale\" content=\"en_US\" \/>\n\t\t<meta property=\"og:site_name\" content=\"Aixia -\" \/>\n\t\t<meta property=\"og:type\" content=\"article\" \/>\n\t\t<meta property=\"og:title\" content=\"LLMOps 2026: From AI Pilot to Production | Aixia\" \/>\n\t\t<meta property=\"og:description\" content=\"Here&#039;s how to take language models from pilot to production with LLMOps. Learn about monitoring, model operations, policy-as-code, and managed MLOps with Aixia.\" \/>\n\t\t<meta property=\"og:url\" content=\"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/\" \/>\n\t\t<meta property=\"article:published_time\" content=\"2026-11-03T09:45:38+00:00\" \/>\n\t\t<meta property=\"article:modified_time\" content=\"2026-11-03T09:45:48+00:00\" \/>\n\t\t<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/aixiasverige\" \/>\n\t\t<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n\t\t<meta name=\"twitter:site\" content=\"@aixiasverige\" \/>\n\t\t<meta name=\"twitter:title\" content=\"LLMOps 2026: From AI Pilot to Production | Aixia\" \/>\n\t\t<meta name=\"twitter:description\" content=\"Here&#039;s how to take language models from pilot to production with LLMOps. Learn about monitoring, model operations, policy-as-code, and managed MLOps with Aixia.\" \/>\n\t\t<meta name=\"twitter:creator\" content=\"@aixiasverige\" \/>\n\t\t<script type=\"application\/ld+json\" class=\"aioseo-schema\">\n\t\t\t{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"BlogPosting\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#blogposting\",\"name\":\"LLMOps 2026: From AI Pilot to Production | Aixia\",\"headline\":\"From Model to Production: LLMOps Will Be the Most Important MLOps Shift of 2026\",\"author\":{\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/author\\\/aixia\\\/#author\"},\"publisher\":{\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/#organization\"},\"image\":{\"@type\":\"ImageObject\",\"url\":\"https:\\\/\\\/aixia.se\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/llmops-produktion-hero.webp?wsr\",\"width\":2048,\"height\":1152},\"datePublished\":\"2026-11-03T10:45:38+01:00\",\"dateModified\":\"2026-11-03T10:45:48+01:00\",\"inLanguage\":\"en-US\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#webpage\"},\"isPartOf\":{\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#webpage\"},\"articleSection\":\"Blog\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#breadcrumblist\",\"itemListElement\":[{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/#listItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/aixia.se\\\/en\\\/\",\"nextItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/category\\\/blog\\\/#listItem\",\"name\":\"Blog\"}},{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/category\\\/blog\\\/#listItem\",\"position\":2,\"name\":\"Blog\",\"item\":\"https:\\\/\\\/aixia.se\\\/en\\\/category\\\/blog\\\/\",\"nextItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#listItem\",\"name\":\"From Model to Production: LLMOps Will Be the Most Important MLOps Shift of 2026\"},\"previousItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/#listItem\",\"name\":\"Home\"}},{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#listItem\",\"position\":3,\"name\":\"From Model to Production: LLMOps Will Be the Most Important MLOps Shift of 2026\",\"previousItem\":{\"@type\":\"ListItem\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/category\\\/blog\\\/#listItem\",\"name\":\"Blog\"}}]},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/#organization\",\"name\":\"Aixia Group AB\",\"description\":\"Aixia levererar AI-infrastruktur, GPU-kluster, MLOps och IT-drift till nordiska f\\u00f6retag. Exklusiv NVIDIA DGX SuperPOD-partner med egen AI-plattform AiQu.\",\"url\":\"https:\\\/\\\/aixia.se\\\/en\\\/\",\"email\":\"info@aixia.se\",\"telephone\":\"+46317620240\",\"foundingDate\":\"2007-01-01\",\"numberOfEmployees\":{\"@type\":\"QuantitativeValue\",\"value\":60},\"logo\":{\"@type\":\"ImageObject\",\"url\":\"https:\\\/\\\/aixia.se\\\/wp-content\\\/uploads\\\/2026\\\/06\\\/aixia-scaled.png\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#organizationLogo\"},\"image\":{\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#organizationLogo\"},\"sameAs\":[\"https:\\\/\\\/www.facebook.com\\\/aixiasverige\",\"https:\\\/\\\/twitter.com\\\/aixiasverige\",\"https:\\\/\\\/www.instagram.com\\\/aixiasverige\\\/\",\"https:\\\/\\\/www.linkedin.com\\\/company\\\/aixia-ab\",\"https:\\\/\\\/sv.wikipedia.org\\\/wiki\\\/Aixia_Group_AB\",\"https:\\\/\\\/www.wikidata.org\\\/entity\\\/Q140274897\",\"https:\\\/\\\/www.nasdaqomxnordic.com\\\/shares\\\/microsite?Instrument=SE0015192065\",\"https:\\\/\\\/aiqu.ai\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/author\\\/aixia\\\/#author\",\"url\":\"https:\\\/\\\/aixia.se\\\/en\\\/author\\\/aixia\\\/\",\"name\":\"Aixia\",\"image\":{\"@type\":\"ImageObject\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#authorImage\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/97c7b1624ea34cae16855ff481655c2d23e3fc872c9911cbd4bc53ae7dfa0c7d?s=96&d=mm&r=g\",\"width\":96,\"height\":96,\"caption\":\"Aixia\"}},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#webpage\",\"url\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/\",\"name\":\"LLMOps 2026: From AI Pilot to Production | Aixia\",\"description\":\"Here's how to take language models from pilot to production with LLMOps. Learn about monitoring, model operations, policy-as-code, and managed MLOps with Aixia.\",\"inLanguage\":\"en-US\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/#website\"},\"breadcrumb\":{\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#breadcrumblist\"},\"author\":{\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/author\\\/aixia\\\/#author\"},\"creator\":{\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/author\\\/aixia\\\/#author\"},\"image\":{\"@type\":\"ImageObject\",\"url\":\"https:\\\/\\\/aixia.se\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/llmops-produktion-hero.webp?wsr\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#mainImage\",\"width\":2048,\"height\":1152},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\\\/#mainImage\"},\"datePublished\":\"2026-11-03T10:45:38+01:00\",\"dateModified\":\"2026-11-03T10:45:48+01:00\"},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/#website\",\"url\":\"https:\\\/\\\/aixia.se\\\/en\\\/\",\"name\":\"AI-infrastruktur, MLOps &amp; IT-l\\u00f6sningar Norden | Aixia\",\"inLanguage\":\"en-US\",\"publisher\":{\"@id\":\"https:\\\/\\\/aixia.se\\\/en\\\/#organization\"}}]}\n\t\t<\/script>\n\t\t<!-- All in One SEO -->\n\n","aioseo_head_json":{"title":"LLMOps 2026: From AI Pilot to Production | Aixia","description":"Here's how to take language models from pilot to production with LLMOps. Learn about monitoring, model operations, policy-as-code, and managed MLOps with Aixia.","canonical_url":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"BlogPosting","@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#blogposting","name":"LLMOps 2026: From AI Pilot to Production | Aixia","headline":"From Model to Production: LLMOps Will Be the Most Important MLOps Shift of 2026","author":{"@id":"https:\/\/aixia.se\/en\/author\/aixia\/#author"},"publisher":{"@id":"https:\/\/aixia.se\/en\/#organization"},"image":{"@type":"ImageObject","url":"https:\/\/aixia.se\/wp-content\/uploads\/2026\/11\/llmops-produktion-hero.webp?wsr","width":2048,"height":1152},"datePublished":"2026-11-03T10:45:38+01:00","dateModified":"2026-11-03T10:45:48+01:00","inLanguage":"en-US","mainEntityOfPage":{"@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#webpage"},"isPartOf":{"@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#webpage"},"articleSection":"Blog"},{"@type":"BreadcrumbList","@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#breadcrumblist","itemListElement":[{"@type":"ListItem","@id":"https:\/\/aixia.se\/en\/#listItem","position":1,"name":"Home","item":"https:\/\/aixia.se\/en\/","nextItem":{"@type":"ListItem","@id":"https:\/\/aixia.se\/en\/category\/blog\/#listItem","name":"Blog"}},{"@type":"ListItem","@id":"https:\/\/aixia.se\/en\/category\/blog\/#listItem","position":2,"name":"Blog","item":"https:\/\/aixia.se\/en\/category\/blog\/","nextItem":{"@type":"ListItem","@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#listItem","name":"From Model to Production: LLMOps Will Be the Most Important MLOps Shift of 2026"},"previousItem":{"@type":"ListItem","@id":"https:\/\/aixia.se\/en\/#listItem","name":"Home"}},{"@type":"ListItem","@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#listItem","position":3,"name":"From Model to Production: LLMOps Will Be the Most Important MLOps Shift of 2026","previousItem":{"@type":"ListItem","@id":"https:\/\/aixia.se\/en\/category\/blog\/#listItem","name":"Blog"}}]},{"@type":"Organization","@id":"https:\/\/aixia.se\/en\/#organization","name":"Aixia Group AB","description":"Aixia levererar AI-infrastruktur, GPU-kluster, MLOps och IT-drift till nordiska f\u00f6retag. Exklusiv NVIDIA DGX SuperPOD-partner med egen AI-plattform AiQu.","url":"https:\/\/aixia.se\/en\/","email":"info@aixia.se","telephone":"+46317620240","foundingDate":"2007-01-01","numberOfEmployees":{"@type":"QuantitativeValue","value":60},"logo":{"@type":"ImageObject","url":"https:\/\/aixia.se\/wp-content\/uploads\/2026\/06\/aixia-scaled.png","@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#organizationLogo"},"image":{"@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#organizationLogo"},"sameAs":["https:\/\/www.facebook.com\/aixiasverige","https:\/\/twitter.com\/aixiasverige","https:\/\/www.instagram.com\/aixiasverige\/","https:\/\/www.linkedin.com\/company\/aixia-ab","https:\/\/sv.wikipedia.org\/wiki\/Aixia_Group_AB","https:\/\/www.wikidata.org\/entity\/Q140274897","https:\/\/www.nasdaqomxnordic.com\/shares\/microsite?Instrument=SE0015192065","https:\/\/aiqu.ai"]},{"@type":"Person","@id":"https:\/\/aixia.se\/en\/author\/aixia\/#author","url":"https:\/\/aixia.se\/en\/author\/aixia\/","name":"Aixia","image":{"@type":"ImageObject","@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#authorImage","url":"https:\/\/secure.gravatar.com\/avatar\/97c7b1624ea34cae16855ff481655c2d23e3fc872c9911cbd4bc53ae7dfa0c7d?s=96&d=mm&r=g","width":96,"height":96,"caption":"Aixia"}},{"@type":"WebPage","@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#webpage","url":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/","name":"LLMOps 2026: From AI Pilot to Production | Aixia","description":"Here's how to take language models from pilot to production with LLMOps. Learn about monitoring, model operations, policy-as-code, and managed MLOps with Aixia.","inLanguage":"en-US","isPartOf":{"@id":"https:\/\/aixia.se\/en\/#website"},"breadcrumb":{"@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#breadcrumblist"},"author":{"@id":"https:\/\/aixia.se\/en\/author\/aixia\/#author"},"creator":{"@id":"https:\/\/aixia.se\/en\/author\/aixia\/#author"},"image":{"@type":"ImageObject","url":"https:\/\/aixia.se\/wp-content\/uploads\/2026\/11\/llmops-produktion-hero.webp?wsr","@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#mainImage","width":2048,"height":1152},"primaryImageOfPage":{"@id":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/#mainImage"},"datePublished":"2026-11-03T10:45:38+01:00","dateModified":"2026-11-03T10:45:48+01:00"},{"@type":"WebSite","@id":"https:\/\/aixia.se\/en\/#website","url":"https:\/\/aixia.se\/en\/","name":"AI-infrastruktur, MLOps &amp; IT-l\u00f6sningar Norden | Aixia","inLanguage":"en-US","publisher":{"@id":"https:\/\/aixia.se\/en\/#organization"}}]},"og:locale":"en_US","og:site_name":"Aixia -","og:type":"article","og:title":"LLMOps 2026: From AI Pilot to Production | Aixia","og:description":"Here's how to take language models from pilot to production with LLMOps. Learn about monitoring, model operations, policy-as-code, and managed MLOps with Aixia.","og:url":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/","article:published_time":"2026-11-03T09:45:38+00:00","article:modified_time":"2026-11-03T09:45:48+00:00","article:publisher":"https:\/\/www.facebook.com\/aixiasverige","twitter:card":"summary_large_image","twitter:site":"@aixiasverige","twitter:title":"LLMOps 2026: From AI Pilot to Production | Aixia","twitter:description":"Here's how to take language models from pilot to production with LLMOps. Learn about monitoring, model operations, policy-as-code, and managed MLOps with Aixia.","twitter:creator":"@aixiasverige"},"aioseo_meta_data":{"post_id":"122615","title":"LLMOps 2026: From AI Pilot to Production | Aixia","description":"Here's how to take language models from pilot to production with LLMOps. Learn about monitoring, model operations, policy-as-code, and managed MLOps with Aixia.","keywords":null,"keyphrases":{"focus":{"keyphrase":"LLMOps"}},"primary_term":null,"canonical_url":null,"og_title":null,"og_description":null,"og_object_type":"default","og_image_type":"default","og_image_url":null,"og_image_width":null,"og_image_height":null,"og_image_custom_url":null,"og_image_custom_fields":null,"og_video":null,"og_custom_url":null,"og_article_section":null,"og_article_tags":null,"twitter_use_og":false,"twitter_card":"default","twitter_image_type":"default","twitter_image_url":null,"twitter_image_custom_url":null,"twitter_image_custom_fields":null,"twitter_title":null,"twitter_description":null,"schema":{"blockGraphs":[],"customGraphs":[],"default":{"data":{"Article":[],"Course":[],"Dataset":[],"FAQPage":[],"Movie":[],"Person":[],"Product":[],"ProductReview":[],"Car":[],"Recipe":[],"Service":[],"SoftwareApplication":[],"WebPage":[]},"graphName":"","isEnabled":true},"graphs":[]},"schema_type":"default","schema_type_options":null,"pillar_content":false,"robots_default":true,"robots_noindex":false,"robots_noarchive":false,"robots_nosnippet":false,"robots_nofollow":false,"robots_noimageindex":false,"robots_noodp":false,"robots_notranslate":false,"robots_max_snippet":null,"robots_max_videopreview":null,"robots_max_imagepreview":"large","priority":null,"frequency":null,"local_seo":null,"breadcrumb_settings":null,"limit_modified_date":false,"ai":null,"created":"2026-11-03 09:45:48","updated":"2026-11-03 09:48:46","focus_keyword":"LLMOps","additional_keywords":null,"truseo_locale":null,"seo_analyzer_scan_date":null},"aioseo_breadcrumb":"<div class=\"aioseo-breadcrumbs\"><span class=\"aioseo-breadcrumb\">\n\t\t\t<a href=\"https:\/\/aixia.se\/en\/\" title=\"Home\">Home<\/a>\n\t\t<\/span><span class=\"aioseo-breadcrumb-separator\">\u00bb<\/span><span class=\"aioseo-breadcrumb\">\n\t\t\t<a href=\"https:\/\/aixia.se\/en\/category\/blog\/\" title=\"Blog\">Blog<\/a>\n\t\t<\/span><span class=\"aioseo-breadcrumb-separator\">\u00bb<\/span><span class=\"aioseo-breadcrumb\">\n\t\t\tFrom Model to Production: LLMOps Will Be the Most Important MLOps Shift of 2026\n\t\t<\/span><\/div>","aioseo_breadcrumb_json":[{"label":"Home","link":"https:\/\/aixia.se\/en\/"},{"label":"Blog","link":"https:\/\/aixia.se\/en\/category\/blog\/"},{"label":"From Model to Production: LLMOps Will Be the Most Important MLOps Shift of 2026","link":"https:\/\/aixia.se\/en\/from-model-to-production-llmops-will-be-the-most-important-mlops-shift-of-2026\/"}],"_links":{"self":[{"href":"https:\/\/aixia.se\/en\/wp-json\/wp\/v2\/posts\/122615","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/aixia.se\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/aixia.se\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/aixia.se\/en\/wp-json\/wp\/v2\/users\/15"}],"replies":[{"embeddable":true,"href":"https:\/\/aixia.se\/en\/wp-json\/wp\/v2\/comments?post=122615"}],"version-history":[{"count":0,"href":"https:\/\/aixia.se\/en\/wp-json\/wp\/v2\/posts\/122615\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/aixia.se\/en\/wp-json\/wp\/v2\/media\/122611"}],"wp:attachment":[{"href":"https:\/\/aixia.se\/en\/wp-json\/wp\/v2\/media?parent=122615"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aixia.se\/en\/wp-json\/wp\/v2\/categories?post=122615"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aixia.se\/en\/wp-json\/wp\/v2\/tags?post=122615"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}