AI Agents as Users
Agents generate 10-100x more volume than humans. Designed for humans = lost revenue.
You built a beautiful interface. Intuitive. User-friendly. Designers love it. Users convert well. Success.
Then AI agents start using your protocol. They generate 10x the volume of human users. But they’re struggling. Your beautiful UI optimized for humans slows them down.
Agent developer reaches out: “Your UI is great for humans, terrible for agents. We’re moving to your competitor. They have proper API, terminal interface, agent-optimized.”
You lose: $5M-20M in annual agent-generated volume. Because you designed only for humans.
Meanwhile, competitor with “ugly” terminal interface capturing all agent volume. Their human UI worse. Their agent experience better.
Result: They’re processing $100M+ monthly volume. 80% from agents. You’re processing $10M. 90% from humans.
Same protocol capability. Different interface strategy. 10x volume difference.
Two protocols. Both technically capable. Different design approaches.
Protocol A: Beautiful human UI only. Mobile-optimized. No terminal. Limited API documentation. 12 months later: $120M annual volume. 10K human users. Some bots (struggling). $600K revenue at 0.5% fees.
Protocol B: Terminal interface prominent. API-first architecture. Agent-optimized documentation. (Human UI exists but secondary) 12 months later: $1.2B annual volume. 5K humans, 500 agents generating 90% of volume. $6M revenue.
10x volume. 10x revenue. Agents made the difference.
The pattern:
Human-only design:
Visual interface required
Click-based actions
Confirmation dialogs
Beautiful but slow
Mobile-optimized
Agent experience: Painful. Must simulate clicks. Slow. Brittle. Many agents won’t integrate.
Agent-optimized design:
API-first (direct protocol access)
JSON responses (machine-readable)
Terminal interface (human traders + agents)
Documentation as UI (agents learn once, execute forever)
Rate limit clarity (agents need predictability)
Agent experience: Fast. Reliable. Easy integration. Agents choose you.
The numbers:
Human-only design:
Volume: $10M monthly (1000 humans × $10K average)
Agent integration: Difficult (3-6 months custom work)
Agent volume: Minimal (agents choose easier platforms)
Revenue: $50K monthly at 0.5% fees ($600K annually)
Agent-optimized design:
Volume: $100M monthly (500 agents × $200K average)
Agent integration: Easy (1-2 weeks standard API)
Agent adoption: High (developers choose you)
Revenue: $500K monthly at 0.5% fees ($6M annually)
Revenue difference: 10x.
Why agents matter:
Volume generation:
Human trader: $10K-100K monthly volume
AI agent: $200K-5M monthly volume
One agent = 20-50 humans in volume
Frequency advantage:
Human: 10-50 trades daily (decision fatigue)
Agent: 1000+ trades daily (no fatigue)
20-100x transaction frequency
Market making:
Humans can’t market make 24/7
Agents never sleep
Liquidity provision = agent dominated
Arbitrage:
Human arbitrage: Minutes to execute
Agent arbitrage: Milliseconds
Speed arbitrage = agent exclusive
What agents need:
API-first architecture:
REST/GraphQL endpoints
WebSocket feeds (real-time)
Rate limits clear (agents plan around them)
Authentication simple (API keys, not OAuth flows)
Machine-readable everything:
JSON responses (not HTML)
Error codes consistent (agents handle errors)
Schema documentation (agents parse structure)
Versioning clear (agents need stability)
Terminal interface option:
Command-line access
Scriptable actions
Batch operations
No GUI required
Documentation as interface:
Complete API docs (agents read once)
Code examples (copy-paste integration)
Rate limits explicit
Edge cases documented
Investment required:
API development:
REST/GraphQL: $100K-200K
WebSocket feeds: $50K-100K
Documentation: $30K-50K
Terminal interface: $50K-100K
Total: $230K-450K
ROI calculation:
Investment: $230K-450K
Agent volume increase: 10x
Revenue increase: $5M-20M annually
ROI: 11-87x first year
When to invest:
Before agents arrive:
Build agent-friendly from start
Cost: $230K-450K
Capture agent market early
After agents struggling:
Retrofit agent optimization
Cost: $450K-700K (migration complexity)
Stop bleeding volume
Never (human-only):
Cost: $0 upfront
Lose: $5M-20M annually to agent-optimized competitors
The mistake:
“Agents can use our human UI. Just simulate clicks.”
Reality: Agents won’t. Too brittle. Too slow. They’ll choose competitor with proper API.
Agents generate 80-90% of volume on agent-optimized platforms. Zero volume on human-only platforms.
Bottom line:
Human-only design: $600K annual revenue. Agents struggle or leave.
Agent-optimized design: $6M annual revenue. Agents choose you.
Investment: $230K-450K. Revenue increase: $5M-20M annually. ROI: 11-87x.
Agents are users. Different users. Need different interface.
Human UI: Visual, slow, mobile-friendly. Agent UI: API, fast, terminal-accessible.
Can’t serve agents with human-only UI. Like serving desktop users mobile-only site.
Most protocols human-only. Missing 80-90% of potential volume.
Agent-optimize now. API-first. Terminal interface. Documentation complete. Capture agent volume.
Or stay human-only. Watch agents choose competitor. Lose 10x revenue.
Future: Agents dominant users. Volume leaders all agent-optimized.
Design for agents now. Or irrelevant later.
Thank you :)
If your project needs design, brand, product, strategy, and leadership,
let’s talk, hi@dragoon [dot] xyz | Follow: 0xDragoon