Z Tag Tag the AI teammate in your channel
One ZAI in the channel: anyone can hand it a task, anyone can follow up, and it gets the job done with your company's systems
Request beta accessNew users by channel
New users by channel
- App stores4,820
- Search3,590
- Community2,610
- Ads1,580
Ads fell sharply; worth reviewing the campaigns next week
64%
of Z.ai staff with access have handed ZAI a task
75%
of tasks are handed off in channels, the rest in DMs
One @ for every kind of work
Sort user questions
Sorts thousands of user questions one by one into an analysis page anyone in your organization can open
Fix bugs, open merge requests
Finds the cause, adds tests, and attaches a before-and-after comparison, ready to merge once an engineer signs off
Build sites and small tools
Internal pages like progress trackers and monitoring dashboards, built in chat and ready to share with colleagues
Useful to every role on the team
Pinpoint root causes
Pulls the data, checks the logs, and reads the code, then hands back an issue report engineering can act on
About an hour from question to issue report in engineering's hands, and someone picked it up the same day
Draft ticket replies
Checks the account, the logs, and the user's app version first, then writes a reply support can send in one click
One support team: more than 280 tickets in three weeks, over half using its draft, and about three-quarters of those sent without a single edit
Find conversion bottlenecks
Breaks down the conversion funnel to find where users get stuck, requests missing data access on the spot, and carries on once it's approved
Once it found the cause, engineering had it write the fix, which was merged into the release branch in under two days
Review merge requests
Paste in a merge request and @ it. It reviews, re-reviews, watches the pipeline, and merges once you confirm
One mobile team ran nearly 100 reviews through it in two weeks, and it merged about 40 of their merge requests
Check policy clauses
Once legal revises a policy, ZAI checks each clause against the code and data and drafts drop-in replacements. Legal makes the final call
Four policies checked in one evening, and the clauses it drafted went into the new versions that shipped
How it works in channels
Shared by everyone
One person kicks it off, and others follow up without another @
Hand it off and move on
It keeps an eye on ongoing work, speaks up only when something changes, and @-mentions someone only when they're needed
Scheduled tasks
Helps before you ask
When someone has a call to make or work to take on, it offers to help and starts only once they agree
Proactive help
Gets to know your team
Say once in the channel who handles what. It notes it in the background and sticks to it in new threads too
Dream
Plugged into your company's systems
Access Bundles
Ready-made setups for common services, and your internal systems connect too
Dedicated sandboxes
One isolated cloud sandbox per channel, which can also connect to your intranet
Deliverables
Report links and sites only your organization can open, with files sent straight back to the chat
- PostgreSQL
- MySQL
- MongoDB
- Redis
- Kubernetes
- SSH
- GitHub
- GitLab
- Jenkins
- Harbor
- Grafana
- Sentry
- Alibaba Cloud
- Tencent Cloud
- Volcano Engine
- Internal systems
- Datadog
- Notion
- Linear
- Atlassian
- Stripe
- Vercel
- Feishu
- S3-compatible storage
Admins set it up once, and everyone else just @-mentions it
Members don't need to install anything or create a new account
Per-chat access
Admins decide which systems and plugins each channel can use
Credential security
Credentials never reach the model, and ZAI can't read them
Instant approvals
Missing access is requested on the spot, and approvers get a DM and sign off in a click
Usage stats
Your organization picks the model provider. Check usage by channel or by person anytime
What people at Z.ai say
For copy changes like this, honestly just have it do them. It'll open the MR itself
Mobile lead If anything's unclear, ask it first. It knows a lot
Product lead, to a new hire The bot's analysis lines up with what I found
Client apps engineering lead Leave the follow-up analysis to it. Have it dig through the tickets and the backend data
Desktop app engineer Just add a bit more info and the bot will look into it on its own
Mobile lead, to QA No need, just have it do it
Product lead