The primary job of any Customer Care and Employee Service team is working towards resolving customer inquiries in a way that satisfies the customer. Speedy resolutions, providing correct information and adding a personalized touch are just three of the methods teams can use to work towards this goal.

But a good support team can only function if the people in that team are happy and get the room to do their job well. Training is one way to make sure they can handle their job. QA and providing them insights in how they do their job is another way to measure what's going on. And the best way to lower the workload and remove pressure? Make sure a big part of their work is automated and handled for them, not by them.

This is where the Zendesk platform shines. Zendesk Guide, or Knowledge as it's called now, is the cornerstone of your ticket deflection strategy. Making information publicly available to customers is an obvious way to remove the "easy" questions from agents' backlog. It gives customers immediate answers, and escalated questions are the more complex, or more valuable kind of interactions.

Over the years Zendesk' help center with articles has evolved into a powerful tool that makes Knowledge available across many channels. The articles you write are used by the Help Center, your AI Agents and Agent Copilot, and are available to agents in the Knowledge Panel of Agent Workspace.

From a content management perspective the platform provides (sub)categories, user segments and restricted articles to make sure customers get the right info presented.
And your team can make sure the content is correct be leveraging article verification rules, publishing processes with approval, change logs and powerful reporting to detect knowledge gaps.

And while many of the features listed here have been available for a while now, the arrival of LLMs and Generative AI gave Zendesk the boost it needed to take the Help Center one step further.

Let's dive in!

Generative Search

Companies that want to do the bare minimum to provide customer support often create a single webpage with a dozen FAQ questions and answers on a single page.

Good companies evolve quickly into a dedicated Help Center with articles categorized in sections. A living platform where content is updated and added regularly, providing a rich and searchable index of knowledge for their customers.

Customers search the Help Center (or use Google...) and land on an article that contains an answer to their questions.

This method works great, but with the arrival of AI, we're actively seeing Google, ChatGPT and Claude changing the way people interact with knowledge. These new AI platforms are (re)training us to expect custom answers to questions, instead of generic articles that contain the info we need.
This method also works great for AI Agents, so why shouldn't we give the same experience for customers visiting your Help Center directly?

Customers don't want to read articles. They want to get an answer to their question.

With Generative Search, now freely available for all customers, you can leverage that same capability on your Zendesk Help Center. By enabling the feature in the Knowledge Admin Center, the search results page will be updated with a new Quick Answer section that will show a custom response to a customers inquiry on top of the classic list of articles.

Customers can immediately read an answer to their inquiry, and if it's helpful, move on without wasting time reading or parsing through a long support article.
And if they do need the details, the full article is just a quick click away.

If you want to know more about this new feature, take a look at my overview article I wrote last year:

Preview of the new generative search for Zendesk Knowledge Base and Agent Workspace

The new generative AI search features for the Zendesk Help Center and Agent Workspace turn search results into Quick Answers with custom generated responses.

Internal NoteThomas Verschoren

Article Summaries

The new Quick Answer page works great for customers searching, but if someone arrives directly on an article page via Google search, by browsing the Help Center or clicking a link provided by an agent, they're still greeted with a full article to read.

This is where the new Article Summaries come in, similar to quick answers, they give customers insight on the usefulness of an article, before they dive in and read the entire article. (I really should add a similar feature to this blog!)

Article Summaries are available to all article that have upgraded to the new article editor. They can be added to an article via the + menu in the toolbar.

If you've got Agent Copilot (~ Zendesk Advanced AI) in your instance, Zendesk will offer to generate a summary for you. If not, you can write your own. Or you can quickly copy the article into ChatGPT, generate a summary, and copy it into the Summary Block.

Summaries appear on top of your Help Center article and are rendered as a zd-summary-block HTML element.

By targeting this in the CSS code of your Help Center theme, you can modify how the summaries are rendered to make them fit your theme and branding.

zd-summary-block {
  border-left: 4px solid #859fa1;
  background: #f3f6f6;
  border-radius: 4px;
  display: block;
  padding: 20px;
  margin-bottom: 10px;
}

I kinda dig these summaries. I see them popup everywhere on Zendesk Support and they're really useful to find out if it's the right article to read.

From an editor perspective I'd love to see a few updates in the future though:

• Bulk update: add a summary to every article in bulk instead of one by one. Although I can understand that by doing it per article you can validate its correctness before publishing.
• Auto-refresh: in a similar vain, if an article is updated, it would be nice if the summary would refresh automatically too.
• Length, complexity, tone of voice: it would be nice if we could adjust the way summaries are phrased.

Writing tools

Summarizations and generated answers are great, but sometimes customers want to dive into the details and they appreciate thoughtful articles on a topic.

And while your customer care team might know the product best, sometimes they might not be the best writers. Similarly, your engineering team can write a perfectly valid article, but they make the content so complex that no regular customer will read that.

This is where the writing tools of the article editor come in. They adjust your team with writing readable content for the Help Center.

Expand

The expand feature is your "we need this content up today and fast!" solution. At times when an incident happens you might want to publish an article quickly.

For example, your team might see an influx of customer inquiries about purchasing tickets after you switched vendors. Someone on the ticketing team sends you this list of steps a customer should take.

While this list is fine, it doesn't really feel like a fun list to read. With a single click of the expand feature, we can turn this into a great support article. It contains all the key elements of the original list, but with additional padding and text that fits your brand.

Simplify

Now, we can also turn this around. Imagine you've got a team member that likes to write really dense and technical articles (👋).

Their content might read like this:

When you click simplify the article will be rewritten in simpler sentences, you get added formatting and the entire thing just gets easier to grasp.

And while the article summary is still there to show the contents at a glance, customers who want to dive into the article now get easier content to read, while still retaining all the technical details you originally provided.

Tone Shift

The final set of writing tools have more to do with aligning your content with your brands' tone of voice.

Your legal team might provide you with a very stern article. But the goal of customer care is to build a good relationship with your customers, not to scare them away.

So let's turn that hefty article from legal into something that's a bit more customer friendly.

On the flip side, if someone from the customer care team writes a very informal article explaining your process, you might want to make it a bit more professional via the make more formal action

I find myself often writing a few bullet points of actions a customer should take. I then expand the article to get an actual article.
I then do a rough editing round, improving the copy and fixing some nuances the AI didn't grasp. And as a final editor round I run it through a make more friendly step to tidy up the entire article.

Translation

Now, we've done the work writing a good article and we've used Zendesk AI to edit the article to fit our company's tone of voice.

But if your company works across countries you probably will need to translate your content to make it accessible to customers that don't speak your native language.

Where AI Agents and Quick Answer can automatically translate your content to fit your customers' needs, translating the full articles used to be a laborious task that's often done by external companies or that one person in the company who happens to speak German.

Zendesk's newly released AI powered translations for the Help Center provide another avenue to make localizing content easier.

AI translation for articles lets you generate translations for articles directly within the article editor. If your help center supports multiple languages, you can create an article in your default language, then use the AI translation feature to automatically translate the article into any language supported by your help center.

The steps here are very similar to the writing tools we showed earlier.

After enabling additional locales for your Help Center, you can select a language you want to add content for. The writing tools menu shows a translation option, which will take the default locale's article content, and offer a translation of the article. This translation will retain formatting, images e.a.

Note, before you can use this feature you need to enable this in the Admin Center > AI > Agent Copilot > Translate knowledge base articles.

Conclusion

So that's it. An overview of all the new AI-powered capabilities of Zendesk Knowledge.

There's one more thing though. And that's the new Knowledge Builder that Zendesk announced at Relate 2025. Knowledge Builder will allow you to easily create a Help Center from scratch, or fill knowledge gaps by turning your Ticket Data into knowledge articles.

I haven't gotten a change to test this one out yet, but it sure looks promising!