When Mikkel Svane and his co-founders, Morten Primdahl and Alexander Aghassipour, founded Zendesk in 2007, they aimed to revolutionize customer service by creating beautifully simple software that enabled businesses to easily interact with their customers and act on their feedback.

Eighteen years later, one could argue that they have been successful. However, Tom Eggemeier, CEO of Zendesk, hopes they will be one of those rare companies that can reinvent an industry for a second time.

However, this time around, they want to start a resolution revolution because, according to Eggemeier, “The only metric that matters in customer service is resolution.”

He’s not wrong.

Who wouldn’t want to be able to resolve their problems faster and easier?

Customers definitely would.

In fact, research finds that most customers in the UK and the US are frustrated with the effort required to find answers to their questions.

Central to Zendesk’s resolution revolution is its new Resolution Platform, which it launched at its recent annual customer event, Relate, held in Las Vegas.

The Zendesk Resolution Platform is built on five core components:

1. An upgraded platform featuring new AI agents that reason, learn, and adapt to even the most complex inquiries, alongside an AI Agent builder that enables companies to create their own customized AI agents tailored to their unique service needs. It also includes enhancements to their Copilot facility, which features the ability to autonomously run key business processes on behalf of human agents.
2. A Knowledge Graph element that consolidates service knowledge and makes it easily accessible by agents, Knowledge Builder which analyzes past tickets, other interactions and key business context documents to identify gaps in a company’s knowledge base as well as a generative AI function to help produce drafts of articles to help fill those gaps. Finally, a new generative AI-powered search function will be available on all Zendesk-powered help centers.
3. The Actions and Integrations component includes Action Builder, a no-code facility that enables Zendesk customers to connect and automate AI and human agent workflows across any system, and App Builder, a no-code solution that helps customers build and develop custom apps within Zendesk using natural language prompts.
4. A Governance and Control element that provides insight and visibility into how AI agents interpret and respond to customer requests.
5. Finally, there is a Measurements and Insight component that includes Custom Quality Assurance (QA), which analyzes and scores all interactions whether human to human or AI to human and provides agent coaching to ensure the best possible outcome, and AI Insights Hub, which provides an overview of how AI is being used, along with recommendations for further automation opportunities.

While all of these five core components are impressive and make sense, the Knowledge Graph component stood out to me.

Why?

Well, for two reasons:

Firstly, customers have long complained about how difficult it is to find answers to their questions and to serve themselves. The research cited above isn’t groundbreaking; it simply provides even more evidence that customers are and remain frustrated.

Secondly, the comprehensive knowledge graph is an idea that has been long in the making.

I remember talking to Adrian McDermott, Zendesk’s CTO, back in 2018 about a new AI-powered knowledge management product they had developed.

At that time, their application was based on Content Cue technology, which uses artificial intelligence and supervised learning to proactively identify gaps in existing knowledge and identifies where new content needs to be developed.

While impressive, one of the big problems it didn’t solve was the actual writing of the new content. And anyone who has been involved with customer service and knowledge base development knows that getting already pressurized agents to write articles is hard and using third-party agencies or other resources takes time, effort, and resources that they often don’t have.

That was more than four years before ChatGPT burst onto the scene.

Now, their new knowledge graph product can, as Eggemeier explains, “analyze all of your articles, all of your tickets, and all of your voice data to identify any gaps in your knowledge base. We then use generative AI to create articles for you to approve, to help you plug those gaps in your knowledge center. Those new articles can then be automatically translated across languages.”

As someone who has been working on this problem for some time, McDermott adds that he is “really excited about Knowledge Builder because I think it will take some of the grind out of knowledge gap identification, creation and maintenance.”

In a world of ever-expanding and changing product portfolios, new service lines, new customers and new geographies, keeping up with the creation of knowledge to help customers help themselves, empower agents and enable AI agents is no small task.

As McDermott points out, “If the AI revolution is the latest industrial revolution, knowledge is the coal.”

I’ve seen a demo of the Knowledge Builder in action. It was impressive and will go a long way toward solving that problem.

This post was originally published on Forbes.com.

Credit: Photo by Alicia Christin Gerald on Unsplash