Our very own Brandon Carlos has been pushing novel ai implementations across the company. Here’s a taste of what he’s been up to!
I recently built an AI agent that helps me quality-assess my frontline team’s conversations with customers. I call him Quinn. Every week, Quinn does the following for me: - Call’s Intercom’s MCP and scans the past week’s worth of eligible customer support conversations - Runs those anonymized conversations through GPT-5 using a QA rubric that I’ve defined and referencing historical anonymized conversations that I’ve deemed are excellent - Scores each conversation on my rubric criteria, and provides detailed, example-based feedback on what was done well and what could be improved - Creates a detailed report for my review, and pings me on Slack to let me know it’s ready Quinn saves me a few hours a week because he’s helping me find the needles in a haystack. I built this agent on Relay.app, and while it’s quickly become my favorite orchestrator, my advice to Operators who are AI-curious is to ignore the tool and first start with the problem. So much of the AI content on LinkedIn is coming from the SaaS startups building these tools. Don’t let their content intimidate, overwhelm or distract you. Start with the problem you want to solve, then force yourself into learn-and-build mode.