- Welcome
-
Getting started 2 min
- Main content
-
How Zendesk optimized its knowledge for automated AI agent resolutions 25 min
- Wrap up
-
Course summary 4 min
-
Share your feedback 5 min
-
Congratulations 1 min
How Zendesk optimized its knowledge for automated AI agent resolutions
See how Zendesk closed the vocabulary gap between how customers ask questions and how articles are written, and lifted its AI agent resolution rate.
In this Zendesk on Zendesk course, you'll follow the Atomic Knowledge experiment our Self-Service team ran: small, customer-worded question-and-answer pairs added to existing Help Center articles in a hidden block. You'll see how we picked high-usage, low-resolution articles, used a custom AI agent to generate the content at scale, connected it to our AI agent Zea, and measured a 3.5 point rise in automated resolution rate to 56.37% across 362 conversations and 14 distinct article topics. You'll leave with a workflow and an action plan you can apply to your own Help Center without rewriting a single article.
Learning objectives
By the end of this course, you will be able to:
- Explain what Atomic Knowledge is and why it matters. You'll describe the vocabulary gap and how small, customer-focused content helps AI agents find the right answer.
- Design an A/B test to prove the impact. You'll pick high-usage, low-resolution articles, create matched versions, and measure the difference.
- Apply the Atomic Knowledge workflow to your own content. You'll use AI to generate customer-style questions and add them in a hidden block, without rewriting your articles.
Audience
This course is intended for all audiences.
Prerequisites
No prior knowledge is required.