The agentic platform for
support teams

Support teams become AI builders, shipping agents that resolve tickets, investigate issues, and automate any workflow behind support.

Anyone can hop on Claude and try to build a chatbot. But to build one that's actually a full system like this? I don't think that's realistic as a one-man job at home.

Trevor Feldmann
Trevor Feldmann
Co-Founder, Vanquish Trader

Managing the support team used to be about managing humans. Now it's about being someone who can manage an AI system.

Matthew Kim
Matthew Kim
Product Manager, Grid

Being able to use the assistant to create things like runbooks, guidelines, and guardrails, and troubleshoot specific runs has made things quite a bit easier.

Jonathan White
Jonathan White
Staff Support Engineer, Automox

We had three support people. Now we have one, and we're handling 30% more tickets than last year.

Aurélio de Pádua Gandra
Aurélio de Pádua Gandra
Founder & CEO, Abbiamo

AI-first support teams build on Duckie

From resolving tickets to investigating issues and triaging escalations, build an agent for any part of the job.

Plan changes agent
When a customer asks to change their plan, update it in Stripe and confirm with them once it's done.
Duckie is working…
Zendesk· Ticket #8842
Upgrade my plan to Pro (monthly)
SolvedPriya S. · handled in 38s
Hi Priya — you're all set! I've moved you to Pro (monthly) and prorated the difference. Your next invoice will…

What teams build on Duckie

From complex financial workflows to onboarding, technical troubleshooting, and compliance — real teams building far beyond ticket deflection.

Now it's becoming more about someone who can manage an AI system.
Ticket resolutionBackend investigation

Automated their most complex financial-services workflows. Duckie walks each member's full ledger, runs branching cancellations, and reads Grid's backend live — resolving the edge cases that used to live only in an operator's memory.

Read case study

What do you want to build?

For data discrepancy issues that look like system bugs, file a Jira ticket to the data-engineering team

Every frontier model,
switch whenever you want

Wired into all your
tools and data

Centralized agent management

Billing agent12.4k runsLive
Escalation agent3.1k runsRolling out · 50%
Investigation agent412 runsIn testing

Orchestrate agents
on one canvas

Triage agent
Needs eng?Ready to reply?
Investigation agent
Resolution agent
The agentic infrastructure

Everything you need to make AI production ready

Point Duckie at your tools and say what should happen — in plain English.

  • Any tool — MCP, APIs, native integrations
  • Workflows written in plain words
  • No engineers required
Youplain English

“When a customer asks to cancel, check their plan in Stripe, offer a retention discount, then update the subscription.”

Connected
ZendeskStripeSlackIntercom
Billing agentbuilt from your runbook
agent runtime · tool-call trace
› plan · resolve cancellation
› stripe.get_subscription(cus_9f2)
‹ active · $49 / mo
› policy.check(refund) ✓
› retry zendesk.reply (attempt 2)
‹ resolved · reply sent
MCP gateway
Plan
Tools
Verify
Reply
test & simulation96% match
Replaying 2,190 real tickets
modesLiveShadowTesting42 agents · one control plane
Policy engineenforced on 100% of agents
Redact PII before any tool call
Refund > $200 → human approval
Never send to external domains
Sandboxed
0 reach a customer until proven safe
Kill-switch armed
Halt any agent instantly
Full trace
Permission scoping

You know how support should work. Now you can build it.