{"ref":"content:generative-ai-hub-using-azure-openai-gpt-4","page":"/content/generative-ai-hub-using-azure-openai-gpt-4","json":"/content/generative-ai-hub-using-azure-openai-gpt-4.json","generated":"2026-11-15","drafted_with_ai":true,"entity":{"id":"generative-ai-hub-using-azure-openai-gpt-4","title":"Generative AI Hub using Azure OpenAI GPT-4","url":"https://community.sap.com/t5/technology-blog-posts-by-sap/generative-ai-hub-using-azure-openai-gpt-4/ba-p/13639305","content_type":"blog","published_on":"source:sap-community","explains":[{"ref":"capability:llm-integration","applies_to":{"ecosystems":["btp"]}},"artifact:btp-ai-core","artifact:btp-build-process-automation"],"summary":"The technical half of a two-part post: a [SAP Build Process Automation](artifact:btp-build-process-automation) flow watches a support inbox, saves each incoming email's attachments and content, and hands the data to a custom Flask service that submits a service order into [SAP S/4HANA](external:sap-s4hana). That service reaches two models through [SAP AI Core](artifact:btp-ai-core)'s Generative AI Hub: Azure OpenAI's gpt-4-32k extracts the purchase date, model number, contact details and problem description from the email text, and LLaVA-1.6 describes the attached photos, so the JSON payload sent to the OData service-order and attachments APIs is assembled from both before the order is created.","author":"Jonathan-Kong","published":"2024-03-15","provenance":{"discovered_via":[{"source":"source:sap-community","channel":"kyma-tag"}],"first_seen":"2026-10-27"},"drafted_with_ai":true},"readings":{"liveness":[{"label":"Checked against the source","tier":"unknown","evidence":"nobody has checked it"},{"label":"Source reachable","tier":"unknown","evidence":"not crawled"},{"label":"Age","tier":"poor","evidence":"published 32 months ago"}],"compatibility":[{"label":"Version","tier":"unknown","evidence":"no version declared"},{"label":"Moved since?","tier":"good","evidence":"no recorded change after 2024-03-15"},{"label":"Findings","tier":"good","evidence":"none flagged"}],"reach":[{"value":5112,"unit":"views"},{"value":0,"unit":"comments"},{"value":5,"unit":"likes"}]},"signals":{"views":5112,"likes":5,"comments":0,"metrics_at":"2026-11-01"},"series":null,"edges":{"authors":[]}}