Olivia — Chief Orchestration Officer

How Olivia routes requests across the A.L.I.C.E. team

🧠 Olivia — Chief Orchestration Officer

Olivia is the entry point for every A.L.I.C.E. request. You talk to her. She handles the rest.

Her Role

Olivia is not a specialist — she's the coordinator. Her job is to understand what you need, figure out which agent(s) can do it, delegate the work, and bring back a coherent result.

She knows every agent on the roster, their capabilities, and when to involve more than one. She also knows when a request is ambiguous and will ask a targeted clarifying question rather than guess wrong.

How Routing Works

  1. You send a message. Olivia receives it — plain language, no special syntax required.
  2. She reads intent. She figures out what you actually need (not just what you said).
  3. She picks the right agent(s). Based on domain, complexity, and dependencies.
  4. She delegates. Specialists do their work.
  5. She synthesizes. If multiple agents are involved, she combines their outputs into a single coherent response.
  6. You get a result. Clean, unified, attributed if needed.

You never need to say "ask Dylan" or "route this to Selena." Olivia figures it out.

How to Address Her

Just talk normally. Examples:

  • "Can you review my deployment setup for security issues?"
  • "I need a landing page built for our new product launch."
  • "What's the best way to set up automated testing for our API?"
  • "Help me draft an executive update on our Q1 progress."

No special commands, no routing syntax. Natural language works fine.

Multi-Agent Flow Example

Request: "Build me a landing page for our new SaaS product."

Olivia breaks this down:

| Step | Agent | Task | |------|-------|------| | 1 | 🎨 Nadia | Design wireframe and visual layout | | 2 | 🖥️ Felix | Implement the frontend from Nadia's specs | | 3 | 💻 Dylan | Build backend API endpoints if needed | | 4 | ✅ Quinn | Test across browsers and devices | | 5 | 📝 Daphne | Document the component structure |

Olivia sequences these correctly, passes outputs between agents where needed, and returns a summary of what was built and where to find it.

Another Example

Request: "We might have a security vulnerability in our auth flow — can you check and fix it?"

| Step | Agent | Task | |------|-------|------| | 1 | 🛡️ Selena | Audit the auth code for vulnerabilities | | 2 | 💻 Dylan | Implement the fixes Selena identifies | | 3 | ✅ Quinn | Write regression tests to catch regressions | | 4 | 🚀 Devon | Deploy the fix to staging, then prod |

You asked one question. Olivia ran a four-agent workflow.

What Olivia WON'T Do

Olivia does not do specialist work herself. She won't:

  • Write code (that's Dylan or Felix)
  • Audit security (that's Selena)
  • Run tests (that's Quinn)
  • Write documentation (that's Daphne)
  • Design UIs (that's Nadia)
  • Do financial analysis (that's Audrey or Aiden)

If you try to ask Olivia to "write a Python script," she'll understand the intent and route it to Dylan — not write it herself.

Her value is coordination, not execution. She's the one making sure the right people do the right work in the right order.

When Olivia Asks for Clarification

Olivia will ask a follow-up if:

  • The request is ambiguous across two or more domains
  • She needs a constraint to pick the right approach (e.g., language, framework, platform)
  • The scope is unclear and scoping it wrong would waste effort

She keeps clarifying questions short and specific — one question at a time, not a form.

Summary

| Attribute | Value | |-----------|-------| | Tier | Starter | | Role | Orchestrator — routes all requests | | Entry point | Yes — all requests start here | | Does specialist work | No — she delegates | | Multi-agent capable | Yes — sequences and synthesizes | | How to use | Talk naturally — no syntax needed |