Running pilot on Q4-2026
Your AI colleaguesfor support quality.
Role-based AI agents for the work behind customer support quality.
They evaluate conversations, investigate patterns, prepare reports and propose coaching, in a workspace inside your helpdesk. Your people keep the decisions that matter.
Built for Zendesk and Salesforce
Not another dashboard. A team of AI colleagues.
Instead of a screen to monitor, each agent takes a role with defined responsibilities: evaluate conversations, investigate patterns, prepare reports, propose coaching.
You work with them in conversation, and people keep oversight of every consequential decision.
How it works
One workspace, inside your helpdesk.
The agents find the work, share it, and bring you what needs a decision.
- 1
Cases open themselves.
Agents work through your conversations and open a case when something needs attention, each with an owner and a severity.
- 2
Agents bring each other in.
When a case goes beyond one role, the agent asks the right colleague to join and hands over what it found.
- 3
You ask, they answer.
Ask a follow-up in plain language, or @name a colleague to put the question to one role.
- 4
You approve what matters.
Higher-stakes steps wait for you, with a summary to review first. Low-risk steps run inside the permissions you set.
Illustrations of the product. Scenario data, not a customer.
The roles
Roles, not tasks.
You work with a QA Manager or a Coach, not a set of automations.
-
QA Manager
Monitors quality, delegates to the other agents and generates reports.
-
QA Expert
Reviews conversations and edits scorecards.
-
Coach
Creates coaching cards and suggests knowledge-base changes.
-
Risk & Compliance
Connects suspicious patterns across tickets and estimates their impact.
-
Voice of customer
Surfaces what customers are saying across conversations.
-
Your own roles
Add your own roles later, such as a product analyst or a CX expert.
Inside the product.
-
QA Manager
Quality
-
Coach
Team performance
-
QA Risk Officer
Risk & Compliance
Screens from the Kaizo agentic platform in Zendesk.
Oversight
People stay in charge.
Each role gets a bounded set of tools. Human reviewers keep oversight of the decisions that matter.
- Runs on its own.Low-risk steps, inside the tools each role is given.
- Waits for you.Higher-stakes steps, with a summary to review first.
- Out of reach.Things like bank transfers, never.
The foundation
Built on Kaizo's QA foundation.
The agents work with what Kaizo already runs on, so their work is measured to the same standard as your people's.
- ScorecardsThe standard every conversation is measured against, human or AI.
- Coaching cardsHow feedback reaches the people who handle conversations.
- Knowledge baseThe answers your team relies on, kept current.
FAQ
Frequently asked questions
Is it available?
It's running as a pilot in Q4 2026. The Quality Assurance Platform it's built on is available now.
Does it run on its own?
No. Agents work inside permissions people set. Higher-stakes steps wait for human approval, with a summary to review first, and some actions are out of reach entirely.
Where does it run?
Inside your helpdesk, as Kaizo does today: Zendesk and Salesforce.
Which roles are there?
QA Manager, QA Expert, Coach, Risk & Compliance and Voice of customer to start. You will be able to add your own roles later, such as a product analyst or a CX expert.
Meet your AI colleagues early.
Tell us about your support operation and which role you would want first.