What is PACE?
Pattern for Agentic Conversational Experience
The PACE Pattern is a user experience architecture where an AI agent guides users to outcomes through conversation rather than traditional navigation. Instead of browsing, filtering, and searching, users engage in dialogue with an intelligent guide that understands intent and surfaces relevant options.
"Don't make users hunt. Let the guide fish for them."
The Two Meanings
PACE operates on two levels simultaneously:
Layer 1: The Framework (What It Is)
P.A.C.E. = Pattern for Agentic Conversational Experience| Letter | Word | Meaning |
|---|---|---|
| P | Pattern | A repeatable architectural approach |
| A | Agentic | AI-driven, autonomous, guide-like |
| C | Conversational | Dialogue over navigation |
| E | Experience | User-centered interaction design |
Layer 2: The Principles (How It Behaves)
P.A.C.E. = Proactive, Adaptive, Contextual, Efficient| Letter | Principle | Implementation |
|---|---|---|
| P | Proactive | Initiate, don't wait. Guide starts the conversation. |
| A | Adaptive | Match the user's level. Technical with devs, simple with beginners. |
| C | Contextual | Remember and reference. Use conversation history and user context. |
| E | Efficient | Concise, actionable, no fluff. Respect user's time. |
Learn more about the recursive acronym →
The Problem
Traditional interfaces assume users know what they want:
User arrives → Browse categories → Filter → Compare → Decide → PurchaseProblems with this model:
| Issue | Impact |
|---|---|
| Cognitive overload | Too many options paralyze decisions |
| Hidden gems | Best-fit products buried in catalogs |
| Context loss | Each page view loses conversation context |
| Generic UX | Same experience for everyone |
| Documentation gap | Users must read before understanding |
The Solution
PACE replaces navigation with conversation:
User arrives → Guide greets → Conversation → Guide recommends → User decidesHow PACE solves this:
| Traditional UX | PACE Pattern |
|---|---|
| Browse 50 products | "What are you fishing for?" |
| Read documentation first | Guide explains as needed |
| Filter, sort, search | Natural language query |
| One-size-fits-all | Adapts to expertise level |
| Static grid | Dynamic conversation |
The Four Components
Every PACE implementation has four core components:
| Component | Purpose | User Question |
|---|---|---|
| Product | AI-guided catalog | "What can you help me with?" |
| About | Context and trust | "Who are you?" |
| Chat | Conversational interface | "I want to ask questions" |
| Executive Summary | Real-time insights | "What have we discussed?" |
Core Principles
Every PACE implementation must embody these four behavioral principles:
1. Proactive
The guide initiates. It doesn't wait for users to figure out what to click.
Examples:
- "Welcome to the pond. What are you fishing for?"
- "I noticed you're interested in MCP servers. Would you like to see our recommendations?"
- "Based on your question, I think these three products might help."
2. Adaptive
The guide matches the user's expertise level and adjusts its communication.
Examples:
- Technical user: "This MCP server implements the stdio transport protocol with JSON-RPC 2.0 messages."
- Beginner: "Think of MCP servers as plugins that let Claude connect to your tools."
3. Contextual
The guide remembers the conversation and references relevant information.
Examples:
- "You mentioned earlier you're using Claude Desktop..."
- "Since you're interested in both StratIQX and PlayIQX, you might want the bundle."
- "Let me circle back to your question about pricing..."
4. Efficient
The guide is concise and actionable. Every response moves the user forward.
Examples:
- ✅ "Here are 3 MCP servers that fit your needs: [list]"
- ❌ "Well, there are many MCP servers available, and choosing the right one depends on various factors..."
Quick Comparison
| Traditional Storefront | PACE Pattern |
|---|---|
| Grid of product cards | Conversational guide |
| Search bar + filters | Natural language |
| Static navigation | Dynamic dialogue |
| Read documentation | Ask questions |
| Browse → Find → Read → Decide | Ask → Guide → Understand → Decide |
Origin Story
The PACE Pattern emerged from designing a storefront for the Cormorant ecosystem — a suite of MCP servers and AI tools. The breakthrough came while observing cormorants at Mill Pond Park in Richmond Hill, Ontario.
"The bird doesn't browse the pond hoping to bump into fish. It dives with intent, adjusts to conditions, and surfaces with exactly what it needs."
The first implementation, MillPond, features a guide named Cormorant who embodies the hunting efficiency of the bird.
Explore the Cormorant Connection →
See It In Action
The best way to understand PACE is to experience it:
- MillPond Storefront — Reference implementation
- Live Demo — Walkthrough with screenshots
- Quick Start Guide — Build your own in 5 minutes
Research & Publications
PACE Pattern v1.0.1 is published on Zenodo:
Next Steps
New to PACE?
- Explore the Semantic Matrix — Understand PACE's unique structure
- Learn the Components — Product, About, Chat, Executive Summary
- Study the Principles — Proactive, Adaptive, Contextual, Efficient
- Build Your First App — Hands-on tutorial
Ready to Build?
- PACE.js Framework — 15KB JavaScript framework
- First App Tutorial — Step-by-step guide
- Examples — Real implementations