Everyone’s talking about AI agents. But making them actually useful in a business context — not just a clever demo — requires more than a chat prompt. You need orchestration: a way to give agents access to real tools, enforce structure on their outputs, and debug them when things go sideways.
An Automated Procurement Crawler
The Challenge:
Our business development team used to start every morning the same way — manually checking a handful of California government procurement portals for new IT and software bids, then eyeballing each one to decide if it was worth pursuing. It took about an hour and was easy to miss things.
We replaced that with an n8n workflow that runs daily at 6 AM:
The Solution:
Discovery via LLM web search. The workflow calls xAI’s Grok model with its built-in web search tool, passing a prompt that targets specific procurement portals, keywords, and geographic focus areas. Grok returns structured bid data — titles, agencies, deadlines, URLs, solicitation numbers — extracted from live web results.
Qualification via a second LLM. Raw leads aren’t useful if 80% of them are irrelevant. Each bid passes through an OpenAI call (GPT-5-mini in JSON mode) that scores it against a three-dimension RFP framework: Scope Fit, Buyer & Process Fit, and Economic & Strategic Fit. A red-flag gate catches immediate disqualifiers (clearance requirements, excluded geographies) before scoring even begins. The output is a classification — Apply, Maybe, or Don’t Apply — plus a short rationale.
Deduplication and task creation. Before creating anything, the workflow fetches existing Wrike tasks and deduplicates by title, URL, and solicitation number. New opportunities land as Wrike tasks with priority badges set by the qualification score and the full analysis embedded in the task description.
The Results
The team now starts their morning with a pre-filtered, pre-qualified list in Wrike instead of a browser full of procurement portal tabs.
How We Built It
At Beezwax, we’ve been building internal AI agents on top of n8n, an open-source workflow automation platform. n8n acts as the control plane: it handles triggers, authentication, tool routing, and LLM orchestration so we can focus on the business logic. Everything runs on a single self-hosted Docker Compose stack — Postgres, Redis, n8n in queue mode, a Caddy reverse proxy — with workflow JSON files version-controlled in Git and deployed via CI/CD.
What We Learned
Agents are only as useful as their integration surface. A brilliant LLM that can’t actually do anything in your business systems is just a chatbot. n8n gives us a way to wire AI into the tools that actually run the business — visually, version-controlled, and without sending credentials to a third-party platform.
If you’re exploring AI automation for your team, we’d love to chat about what’s possible.