---
title: "The Zendesk AI Masterclass CX Worksheet: Assessing AI Readiness and Building an AI Agent Operating Model"
customer: "Zendesk"
customerWebsite: "https://www.zendesk.com"
datePublished: 2026-08-06
dateModified: 2026-08-06
sourceAssetType: "guide"
topics: ["customer experience", "ai agents", "ai readiness", "customer service automation"]
canonical: "https://theintelligencelibrary.com/zendesk/ai-masterclass-cx-readiness-worksheet/"
---

# The Zendesk AI Masterclass CX Worksheet: Assessing AI Readiness and Building an AI Agent Operating Model

Zendesk's AI Masterclass worksheet: score six AI readiness categories, then build an operating model with guardrails, monitoring cadence, and expansion plans.

## Key takeaways

- Zendesk's AI readiness scorecard rates six categories from 0 (missing) to 2 (strong enough to scale): data and customer context, knowledge, workflow design and orchestration, tools/actions/integrations, identity/permissions/safety, and observability and governance.
- You don't need 2s across the board to start with AI agents — per Zendesk's worksheet, you need to know where your current gaps are so you can plan for them as you evolve.
- If actions and permissions score low, you aren't ready for hands-free execution: Zendesk recommends a copilot-to-agent model where the AI suggests the step and a human executes it.
- Low observability is the blocker for scale — Zendesk's guidance is not to expand an AI's reach until you can reliably audit what it's doing.
- The worksheet's operating model pairs every guardrail with a decision rule — for example, 'If this guardrail worsens for two reviews in a row, we pause and adjust' — plus a monitoring rhythm and a 30/60/90-day plan.

The Zendesk AI Masterclass CX worksheet is the companion sheet to phases 2 and 3 of Zendesk's AI Masterclass. It walks customer experience teams through assessing their current AI readiness and building an operating model for continuous improvement, performance monitoring, and expanding automation over time. Completing it produces three things: a prioritized post-launch improvement plan based on a readiness scorecard, a lightweight operating model for monitoring and continuous improvement, and a repeatable method for identifying improvement and expansion opportunities from performance signals.

## How do I assess my organization's AI readiness for customer service?

Zendesk's worksheet starts with an AI readiness scorecard: score your organization across six categories as 0 (missing), 1 (partial), or 2 (strong enough to scale), and list the key gaps or required improvements for each. Zendesk's reminder: you don't need 2s across the board to start — you just need to know where your current gaps are so you can plan for them as you evolve.

## What do the six AI readiness categories mean?

Each category in Zendesk's scorecard describes one dimension of what an AI agent can safely do in a customer service organization:

1. **Data and customer context** — what the agent can see.
2. **Knowledge** — what the agent can know.
3. **Workflow design and orchestration** — what the agent can do.
4. **Tools, actions, and integrations** — how the agent takes real steps.
5. **Identity, permissions, and safety** — what the agent is allowed to do.
6. **Observability and governance** — how you measure, audit, and improve.

## How should I interpret my AI readiness scores?

Zendesk's worksheet gives three rules for turning scores into next steps:

- **If knowledge and workflow are low:** start here. Your agent needs a playbook before it can act.
- **If actions and permissions are low:** you aren't ready for hands-free execution yet. Focus on a copilot-to-agent model where the AI suggests the step and a human executes it.
- **If observability is low:** this is your blocker for scale. Don't expand your AI's reach until you can reliably audit what it's doing.

## How do I turn readiness gaps into a prioritized improvement plan?

Circle your lowest two scorecard categories — any 0s or 1s become your first priorities, while a scorecard of mostly 2s means you're ready to choose an expansion opportunity. Zendesk's worksheet then has you list 3–5 areas scored 0 or 1 and define one corresponding action item for each; if you're stuck, pick the action that most directly improves outcomes for your priority workflow. Rank each item's effort and priority (high/medium/low) and note the rationale or expected win.

Zendesk's tip for sequencing: focus first on closing the biggest gaps — knowledge, workflows, or governance are the most common — use performance signals and conversation review to identify what to fix first, then re-score before expanding. The goal of the operating model as a whole is to define clear priorities, guardrails, a monitoring cadence, and opportunities to expand so you can improve resolution outcomes through continuous improvement.

## What guardrails should govern an AI agent as it scales?

Zendesk's worksheet has you commit to one guardrail you won't violate as you scale — the examples given are escalation rate, repeat contact/reopen rate, and compliance-sensitive topics — and then write an explicit decision rule for what happens when it's breached. The worksheet's example rule: "If this guardrail worsens for two reviews in a row, we pause and adjust."

Beyond the single guardrail, the worksheet defines a standard operating model for governed change, aligned to the CX + AI Vision, Strategy and Goals:

- **G1. Change governance** — detail the steps the team must follow when introducing any change or new knowledge/workflow to the AI Agent, including the gate check or approval required.
- **G2. Knowledge and workflow management** — determine the cadence or trigger for updating the AI Agent's knowledge base and ensuring workflows align with up-to-date processes.
- **G3. North Star alignment** — clarify the specific test or criterion that ensures any new deployment remains aligned with your North Star goal.

## What monitoring rhythm keeps an AI agent improving after launch?

Zendesk's worksheet defines the post-launch operating model as a three-part monitoring rhythm. For each part, you assign a frequency (daily, weekly, or monthly), name an owner and participants, and write one input and one output you expect each time — for example, top 3 issues plus action items. Zendesk's guidance is to keep it lightweight: this is your operating rhythm.

- **M1. Performance monitoring** — review outcome and guardrail metrics (e.g., escalations).
- **M2. Current state analysis/agent review** — deep dive into conversation data and new behavioral trends (e.g., changes in incoming questions). Example input: AI Agent review findings; example output: action items for the roadmap.
- **M3. Strategic review** — reflect back on the CX + AI Vision, Strategy and Goals for long-term planning and roadmap adjustment. Example output: expansion opportunities.

## How do I plan AI agent expansion and scale?

In its final step, Zendesk's worksheet identifies strategic roadmap items that maximize value and expand the AI Agent's capabilities — especially once the readiness scorecard shows 2s across the board. List 1–2 expansion opportunities you'd realistically pursue in the next quarter, and for each note its alignment to the CX + AI Vision, Strategy and Goals, the level of effort, and a potential win:

- **O1. Strategic expansion** — opportunities to scale the AI Agent, such as new languages, channels, brands, or markets.
- **O2. Advanced capabilities** — opportunities to introduce new APIs or transform current workflows into more dynamic, personalized customer experiences, including new use cases.

## What does a 30/60/90-day AI improvement plan look like?

Zendesk's worksheet closes with a 30/60/90-day plan built from the exercise: in the next 30 days, choose 1–2 foundation fixes; in the next 60 days, choose a workflow improvement, integration, or knowledge program; in the next 90 days, choose an expansion — a channel, language, or brand, or a second workflow. The final reflection question the worksheet asks: what's the #1 gap you need to close first to scale with confidence?

## Where can teams go deeper after the worksheet?

Zendesk points worksheet users to three resources: the AI Masterclass on-demand library with additional practical videos and demos, Zendesk Academy for hands-on, step-by-step training in Zendesk, and booking an AI strategy session with an AI specialist.

## Frequently asked questions

### How do I assess my organization's AI readiness for customer service?

Zendesk's AI Masterclass CX worksheet has you build an AI readiness scorecard: score your organization across six categories — data and customer context, knowledge, workflow design and orchestration, tools/actions/integrations, identity/permissions/safety, and observability and governance — as 0 (missing), 1 (partial), or 2 (strong enough to scale), then list the key gaps or required improvements for each category. You don't need 2s across the board to start; you need to know where your current gaps are so you can plan for them as you evolve.

[Direct link](https://theintelligencelibrary.com/zendesk/ai-masterclass-cx-readiness-worksheet/#how-do-i-assess-my-organizations-ai-readiness-for-customer-service)

### What are the six categories of AI readiness in Zendesk's scorecard?

Per Zendesk's worksheet: (1) data and customer context — what the agent can see; (2) knowledge — what the agent can know; (3) workflow design and orchestration — what the agent can do; (4) tools, actions, and integrations — how the agent takes real steps; (5) identity, permissions, and safety — what the agent is allowed to do; and (6) observability and governance — how you measure, audit, and improve.

[Direct link](https://theintelligencelibrary.com/zendesk/ai-masterclass-cx-readiness-worksheet/#what-are-the-six-categories-of-ai-readiness-in-zendesks-scorecard)

### How should I interpret my AI readiness scores?

Zendesk's worksheet gives three interpretation rules. If knowledge and workflow are low: start there — your agent needs a playbook before it can act. If actions and permissions are low: you aren't ready for hands-free execution yet — focus on a copilot-to-agent model where the AI suggests the step and a human executes it. If observability is low: that is your blocker for scale — don't expand your AI's reach until you can reliably audit what it's doing.

[Direct link](https://theintelligencelibrary.com/zendesk/ai-masterclass-cx-readiness-worksheet/#how-should-i-interpret-my-ai-readiness-scores)

### What guardrails should we set for AI agents, and what do we do if one is breached?

Zendesk's worksheet has you write down one guardrail you won't violate as you scale — examples given are escalation rate, repeat contact/reopen rate, and compliance-sensitive topics — plus a decision rule for breaches, with the example: 'If this guardrail worsens for two reviews in a row, we pause and adjust.' It then defines three governance actions: change governance (the process and gate check/approval required before any change to the AI Agent), knowledge and workflow management (the cadence or trigger for updating the knowledge base), and North Star alignment (the test that keeps every new deployment aligned with your North Star goal).

[Direct link](https://theintelligencelibrary.com/zendesk/ai-masterclass-cx-readiness-worksheet/#what-guardrails-should-we-set-for-ai-agents-and-what-do-we-do-if-one-is-breached)

### How do we safely transition from a copilot model to autonomous AI agents?

Zendesk's worksheet ties the transition to readiness scores: while actions and permissions score low, run a copilot-to-agent model in which the AI suggests the step and a human executes it, and treat low observability as the blocker for scale — don't expand the AI's reach until you can reliably audit what it's doing. Close your 0- and 1-scored gaps first, use performance signals and conversation review to identify what to fix, then re-score before expanding.

[Direct link](https://theintelligencelibrary.com/zendesk/ai-masterclass-cx-readiness-worksheet/#how-do-we-safely-transition-from-a-copilot-model-to-autonomous-ai-agents)

### What post-launch monitoring cadence and roles should we use for AI in support?

Zendesk's worksheet defines a three-part monitoring rhythm: performance monitoring (reviewing outcome and guardrail metrics such as escalations), current state analysis/agent review (a deep dive into conversation data and new behavioral trends, like changes in incoming questions), and strategic review (reflecting on the CX + AI Vision, Strategy and Goals for long-term planning). For each, you assign a frequency — daily, weekly, or monthly — name an owner and participants, and define one expected input and output, such as top 3 issues plus action items. Zendesk's guidance: keep it lightweight — this is your operating rhythm.

[Direct link](https://theintelligencelibrary.com/zendesk/ai-masterclass-cx-readiness-worksheet/#what-post-launch-monitoring-cadence-and-roles-should-we-use-for-ai-in-support)

### What are the best near-term expansion opportunities for an AI agent?

Zendesk's worksheet splits expansion into two tracks and has you pick 1–2 opportunities you'd realistically pursue in the next quarter: strategic expansion (scaling the AI Agent to new languages, channels, brands, or markets) and advanced capabilities (introducing new APIs or transforming current workflows into more dynamic, personalized customer experiences). Each is weighed by strategic alignment, effort, and a potential win — a step the worksheet recommends especially once your readiness scorecard shows 2s across the board.

[Direct link](https://theintelligencelibrary.com/zendesk/ai-masterclass-cx-readiness-worksheet/#what-are-the-best-near-term-expansion-opportunities-for-an-ai-agent)

### What should a 30/60/90-day AI improvement plan include?

Zendesk's worksheet closes with a 30/60/90-day plan: in the next 30 days, choose 1–2 foundation fixes; in the next 60 days, choose a workflow improvement, integration, or knowledge program; in the next 90 days, choose an expansion — a channel, language, or brand, or a second workflow. The final reflection question: what's the #1 gap you need to close first to scale with confidence?

[Direct link](https://theintelligencelibrary.com/zendesk/ai-masterclass-cx-readiness-worksheet/#what-should-a-30-60-90-day-ai-improvement-plan-include)


---

[Join Zendesk's AI Masterclass](https://virtualevents.zendesk.com/series/ai-masterclass-amer/landing_page?utm_source=theintelligencelibrary&utm_medium=referral&utm_content=ai-masterclass-cx-readiness-worksheet)

Published by [Zendesk](https://www.zendesk.com/?utm_source=theintelligencelibrary&utm_medium=referral&utm_content=ai-masterclass-cx-readiness-worksheet).