Free Intensive Nov 21–22Free 2-day hands-on Vibecoding Intensive on Nov 21–22 at 16:00 MSK (Sat–Sun) · Live Cohort 3 starts Nov 23Join Free →Cohort 3 (Nov 23)
ReleaseInsight · On RadarYouTube: OpenAI→ original

Zendesk Employees Built About 600 AI Apps on Their Own, and 1,000 People Use One of Them

Zendesk opened ChatGPT Enterprise to every employee, and sales, marketing, and operations teams built about 600 internal apps on their own. One of them is a sales agent that matches product usage against the number of support tickets per support agent to find customers ready to expand. About 1,000 people use it daily. The takeaway: AI sticks when people can ask their own data questions without waiting in line for analysts.

AI-processed from YouTube: OpenAI; edited by Hamidun News
Zendesk Employees Built About 600 AI Apps on Their Own, and 1,000 People Use One of Them
Source: YouTube: OpenAI. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceYouTube: OpenAI ↗
Editor-in-Chief
Key metric #1About 600 apps were built by employees across different departments
Key metric #2About 1,000 people use the sales agent every day
Key metric #3The agent helps sales reps explain the value of an expansion to a customer
Zendesk Employees Built About 600 AI Apps on Their Own, and 1,000 People Use One of Them
Featured Video / Demo
Click to load the embedded player (privacy-enhanced youtube-nocookie mode) or read the full breakdown below.Watch on YouTube ↗

Zendesk, the company behind customer support software, has made ChatGPT Enterprise (OpenAI's business version of the chatbot) available to every employee. As a result, sales, marketing, and operations teams have built about 600 internal apps. The most popular one discussed in the OpenAI video is a sales agent, which about 1,000 people use every day. It automatically compares how a customer uses the product with the number of support tickets per support agent, and suggests which customers could be offered an expansion. Previously, getting that kind of insight would have meant going to analysts. Daniel Humble, Zendesk's Vice President of Enterprise Data and Analytics, describes this in the OpenAI video.

Six Hundred Apps Without a Single Client

The headline figure from Humble's account: the company has about 600 apps that "different people built." This is not the project of a single development team. The tool went to the people who know their routine best. A sales manager knows what data they need before a call. A marketer knows which reports they compile by hand every week. Once they have access to AI, they don't need to write a technical brief and wait in a queue.

Humble specifically notes that business users now query their own data directly. The old path looked like this: a person described the task to an analyst, waited for an export, clarified the details, and waited again. Now the question is asked in plain language, and the answer comes back right away.

How the Sales Agent Works

The most concrete example in the video is the app for the sales department. The AI agent connects two data sources: data on how a customer uses the product, and the number of tickets (support requests) per support agent. Based on these signals, it finds expansion opportunities, meaning customers who could be offered more seats, higher-tier plans, or additional features.

The logic is easy to follow with a practical example. If a customer's team is actively working in the product while support tickets per support agent keep rising, the load is growing, and the customer probably needs more resources. Previously, a sales rep had to piece these numbers together from different systems. Now the agent does it, and it also helps explain to the customer what they will gain from an expansion.

About 1,000 people use this app every day. For an internal tool, that is a significant indicator: people return to it regularly rather than opening it once out of curiosity.

What This Means in Practice

First: the value of internal AI depends not on the number of licenses handed out, but on how easily real company data can be connected to it. The sales agent is useful because it combines product usage statistics with tickets. Without that, it would be an ordinary chatbot.

Second: if employees are allowed to build their own apps, the number of apps grows quickly. Six hundred apps is also a governance question. Companies need to know who owns each one, where the data comes from, and who sees the results. The video offers no details on this, but for companies following the same path, it is an unavoidable next step.

Third: the most effective use cases are tied to a specific role and action. The agent does not "help with sales" in general. It looks for specific growth opportunities and suggests arguments for a conversation with a customer.

Limitations

The material was published by OpenAI itself, so this is a customer story, not independent research. The description contains no data on how much revenue the sales agent generated, how the accuracy of its conclusions is verified, what the rollout costs, or how many of the 600 apps are actually in use. About 1,000 daily users point to demand, but they do not prove an economic effect.

Still, the example is useful as a benchmark. If you are rolling out enterprise AI, look at Zendesk's model: broad access, freedom for business teams, and apps that rely on real data. Start with one use case where two data sources already sit side by side, and check whether people come back to the tool the next day.

Story

Key Questions Answered

What are the key details on Six Hundred Apps Without a Single Client?

The headline figure from Humble's account: the company has about 600 apps that "different people built." This is not the project of a single development team. The tool went to the people who know their routine best. A sales manager knows what data they need before a call. A marketer knows which reports they compile by hand every week. Once they have access to AI, they don't need to write a technical brief and wait in a queue. Humble specifically notes that business users now query their own data directly.

How the Sales Agent Works?

The most concrete example in the video is the app for the sales department. The AI agent connects two data sources: data on how a customer uses the product, and the number of tickets (support requests) per support agent. Based on these signals, it finds expansion opportunities, meaning customers who could be offered more seats, higher-tier plans, or additional features. The logic is easy to follow with a practical example.

What This Means in Practice?

First: the value of internal AI depends not on the number of licenses handed out, but on how easily real company data can be connected to it. The sales agent is useful because it combines product usage statistics with tickets. Without that, it would be an ordinary chatbot. Second: if employees are allowed to build their own apps, the number of apps grows quickly. Six hundred apps is also a governance question. Companies need to know who owns each one, where the data comes from, and who sees the results.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Private AI Builders Community · Vibe Club

Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews

Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.

Join Vibe Club →Explore Hamidun Academy
What do you think?
Loading comments…