N8n-orchestrated pilot workflow

n8n Automation / AI Workflow Pilot We are looking for an experienced n8n developer to help build a pilot workflow that automates research, summarisation, and structured output generation from a small set of user-provided inputs. The aim is to create a working proof of concept that can take basic entity details, gather relevant publicly available information from defined sources, use an LLM to generate a structured briefing, and then pass the output through a human review step before any approved information is saved or pushed into another system. This was originally considered as a custom-coded application, but we are now exploring n8n as a faster way to validate the workflow, test the process with real examples, and iterate before deciding whether a bespoke build is needed later. What we need We need someone who can help design and build an n8n-based workflow that includes: A simple input trigger or form for capturing basic details. API-based lookups from selected public data sources. Web/search-based research steps where appropriate. LLM integration, likely via Claude or another major model provider. Structured output generation, ideally in JSON or another format suitable for review and onward use. Source capture, confidence/uncertainty indicators, and basic audit logging. A human review stage where generated fields can be checked, edited, approved, or rejected. A final handoff step into a CRM, database, spreadsheet, or other staging system. Error handling, retries, logging, and clear documentation of the workflow. Important requirements The workflow should produce structured, reviewable outputs with clear evidence or source links where possible. Some fields may come from reliable public sources, while others may be inferred by the LLM, so the workflow must clearly distinguish between verified information and generated/inferred content. Skills we are looking for The ideal freelancer will have experience with: n8n workflow design and implementation. HTTP/API integrations. LLM integrations, especially Claude, OpenAI, or similar. Structured prompt design and JSON outputs. Data validation and error handling. Webhook-based workflows. Building workflows that are maintainable and well documented. Initial engagement We would like to start with a short exploratory call to walk through the existing specification, discuss the proposed n8n approach, identify any gaps or assumptions, and agree the best pilot architecture.