When we first built Autonomos, every agent had a boring name. "Lead Scanner." "Quality Reviewer." "Content Writer." Descriptive. Forgettable.
Then we switched to Game of Thrones characters and something unexpected happened: people actually understood how the system worked.
The problem with "AI Assistant"
Most AI products describe their agents the same way: "Our AI assistant analyzes your data and provides insights." Cool. That could be literally anything.
When everything is a generic assistant, you can't build a mental model of what's happening. Who did what? Where does quality come from? Why did this lead get rejected? Black box.
Named characters fix this instantly. When you see that Jon Snow rejected 224 of 247 prospects — you understand something about the standard. The character tells you the personality of the filter. That's not something "Lead Quality Module" communicates.
Each name was chosen for a reason
Littlefinger finds things. He always knew where the leverage was. Our Littlefinger scans hundreds of companies and finds every one that fits. He doesn't miss.
Jon Snow sets standards. His word is final. No appeals, no explanation. As a lead qualifier, he cuts everything that doesn't meet the bar.
Tyrion writes. Sardonic, particular, pleased when work is good. He'd be offended if his emails sounded like cold outreach.
Daenerys judges. Cold approval. Rejected means rejected. Her 80% pass rate means the 20% that got sent back deserved it.
Jaqen H'ghar delivers. Zero emotion. Task confirmed. Nothing late. Everything on schedule. Always.
The thing we didn't expect
People talk about the characters. When someone says "Daenerys rejected 6 of my scripts and honestly she was right" — that's a referral engine you can't manufacture.
The names give people specific language to describe what happened. And specific stories beat generic claims every time.
It's not just branding
Every agent description is written in character. Jon Snow doesn't explain himself. Littlefinger implies he was already done before you asked. The Hound states what happened, no decoration.
This consistency makes the system feel like a real team with real personalities, not "AI" doing "things." That's the difference between a tool and something people actually remember.