Facilitator Guide: AI Search & Digital Assistants
You are a digital agency helping a client deploy ai12z's AI Search and Digital Assistant experience on their own site. This guide is how you scope, deploy, and grow that engagement.
This is a sibling practice to Get Found in AI Search, ai12z's agency program for AI discoverability (getting a brand found, cited, and recommended by ChatGPT, Gemini, and Perplexity). That program answers "how do we get found by AI." This one answers the next question a client asks once they are found: "now what happens when that visitor actually lands on our site?" ai12z calls the two together an AI Experience Platform — discovery, then engagement, in one connected system. If a client is asking both questions, run both programs; they share a platform and a philosophy, not just a vendor.
Who This Is For
Digital agencies and consultants deploying ai12z's AI Search and Digital Assistant capabilities for their own clients — everything from a lightweight AI answer box added to an existing search page, up to a full agentic assistant wired into CRM, commerce, and support systems. If you're looking for ai12z's own technical reference documentation for a specific capability, this guide links out to it throughout rather than duplicating it — start with Custom Integration Overview if you already know exactly what you're looking for.
The Operating Model: Scope → Deploy → Optimize
| Verb | What It Means | What It Looks Like |
|---|---|---|
| Scope | Understand the client's current search/support stack, CMS, integration needs, and industry constraints, then pick a starting tier | The Discovery & Scoping intake, resulting in a Crawl, Walk, or Run engagement |
| Deploy | Ship the chosen surface — from a Knowledge Box layered on existing search up to a fully agentic Bot — against real, ingested content | A configured, tested, branded deployment reviewed in Your Working Session |
| Optimize | Use Logs, Analyze, and Test Plans to improve answer quality, then graduate the client to the next tier as trust and value are proven | Trust & Governance and the tier-graduation path in Engagement Tiers |
Unlike a one-time website project, this is designed to be a relationship that grows — a client who starts at Crawl with a simple Knowledge Box is a client who can graduate to Walk and Run once the first deployment proves its value. See The Agency Service Model for how that growth path becomes recurring revenue.
Format at a Glance
| Starting point | Determined by Discovery & Scoping — not every client needs the same tier on day one |
| Engagement tiers | Crawl, Walk, Run — see Engagement Tiers for the full capability matrix |
| Live session length | 45–60 minutes for a scoping/demo session — see The Agenda |
| Output | A scoped tier recommendation, a working deployment plan, and a client-ready walkthrough of what will be built |
Guide Chapters
Chapters 1–8 are written to be read in order — they walk you through scoping and running the first client conversation. Chapters 9–16 are reference material, meant to be consulted as needed rather than read start to finish.
Read in Order
| # | Chapter | Covers |
|---|---|---|
| 1 | Facilitator Guide Overview (this page) | The operating model, who this is for |
| 2 | Why This Matters Now | The discovery-to-engagement narrative; bridging from the GEO/AEO program |
| 3 | The ai12z Advantage | RAG + ReAct + MCP, and why that's more than a chatbot |
| 4 | Engagement Tiers | Crawl / Walk / Run, and the capability matrix that keeps tiering honest |
| 5 | Discovery & Scoping | The intake that determines a client's tier, plus guardrails for live demos |
| 6 | The Agenda | Run of show for the scoping/demo session |
| 7 | Deployment Surfaces | Knowledge Box vs. Modal vs. Search Results Page vs. Bot vs. APIs |
| 8 | Your Working Session | The actual client-facing walkthrough |
Reference Material
| # | Chapter | Covers |
|---|---|---|
| 9 | Knowledge Foundation | Connectors, web scraping, Image AI, auto-create agent |
| 10 | Guided Experiences | Welcome screens, Directives, Landing Page Experience Controls, Vibe Coding |
| 11 | Integrations & Actions | ReAct, OOTB + custom integrations, MCP, Forms, Commerce Assistant, Geolocation |
| 12 | Service Operations | Salesforce/Zendesk Knowledge & Cases, Live Agent Escalation, Activepieces |
| 13 | Trust & Governance | Content filtering, Logs/Analyze, Test Plans, controlled/compliance-first answer modes |
| 14 | The ai12z Toolbox | Every capability in this guide, mapped to tier and existing docs |
| 15 | Managing Multiple Clients | Sub-organizations, billing, branding, environment promotion |
| 16 | The Agency Service Model | Sellable packages mapped to Crawl/Walk/Run |
| 17 | Close & Next Steps | Wrap-up checklist |
Related Documentation
- Get Found in AI Search: Agency Program — The sibling program for AI discoverability, the "get found" half of the same platform story
- AI Assistant Agency Playbook: Program Overview — Back to the program landing page
- Custom Integration Overview — Full technical reference for how integrations and response handling work under the hood