Skip to main content
AI-powered support

The support stack, rebuilt around the agent.

One queue across email, chat, and WhatsApp. Agents that classify, look things up, act, and escalate when policy says so.

Grounded in your docs · gated by your policy · audited end-to-end·Security overview →
S
Subport agent
online · demo transcript
connecting to agent…
Native integrations
and APIs for the long tail
Why teams switch

Support tools weren't built for agents that act

Most teams are stuck between legacy ticketing and bolt-on assistants that never close the loop. Subport closes it.

More seats, same backlog

More seats logged in doesn't mean more tickets resolved. Subport puts automation where it deflects, resolves, and hands off cleanly.

AI that stops at a draft

Copilots suggest text. Customers still wait. Subport takes the action end-to-end, with gates when risk is high.

Weeks to wire channels and playbooks

You shouldn't need a migration project to get value. Connect channels, attach knowledge, and route work in hours — not quarters.

The platform

Six surfaces. One mental model.

Click around. These are interactive demos, not screenshots.

Inbox

One queue. Every channel.

Email, chat, voice, WhatsApp, X, Instagram, SMS — same thread, same agents, same shortcuts.

All 142Mine 12Unassigned 38
P1Sara Chen2m
Can't log in after SSO change
agent
VOICEMarcus Lee8m
Voicemail · device pairing failed
human
SOC@miradev12m
X mention · export to CSV broken
agent
CHATAnonymous18m
Live chat · upgrade to team plan?
agent
WA+91 98765…34m
WhatsApp · package delayed 4d
agent
P3Aria Patel1h
GDPR data export request
agent
Macros

Saved actions, one keystroke.

Codify your team's playbook as named actions with parameters. People or agents call them by name.

⌘1
reset_sso(user_email)
Force SSO re-auth + clear cached tokens
tool
⌘2
escalate_to_eng(thread_id, severity)
Page on-call with thread + repro steps
human
⌘3
wism_followup(order_id)
Where is my order — auto-fetch carrier
agent
⌘4
gdpr_export(customer_email)
Generate full data export · audit-logged
tool
⌘5
upgrade_plan(workspace_id, tier)
Propose plan change + book a call
human
⌘6
close_with_csat()
Resolve + send CSAT survey + tag
agent
Channels

Add channels in <5 minutes.

Each one feels native. None spawn a new tab.

Command palette

Every action, one keystroke away.

Tickets, macros, jumps. Fuzzy match across everything your team does.

Tickets
Reset SSO for sara@acme · TKT-4829
Reset password · mobile app · TKT-4803
Macros
reset_sso⌘1
reset_password
Navigate
Go to Knowledge → SSO setupG K
↑↓ navigate· openSubport
Insights

What got resolved, and why.

Auto-resolution rate, intent volume, agent-vs-human split.

Auto-resolve · 12w74%+18pt
order_status96%
login_issue89%
account_change71%
billing_question84%
Agent runs

Every tool call, traceable.

Open any conversation, see exactly what the agent decided, what tools it called, what humans approved.

Run · sso-reset#4829
RUNNING12.8s · 5 tools · 1 human
The conviction

Built for operators and engineers.

Ship macros and routing without filing an IT ticket. Extend with code when the workflow gets sharp.

01

Operators ship in an hour.

Connect a channel, attach a knowledge base, define gates for risky actions. The repetitive tier takes a coffee break.

02

Engineers don't hit a wall.

APIs, webhooks, custom tools, and workflows that match how your org actually resolves work — not a toy canvas.

03

Seats for people. AI that scales with the plan.

Pick a tier sized to your team and automation volume. No surprise per-message tax — current numbers on the pricing page.

What teams optimize for

Targets that match how support is measured.

Illustrative benchmarks for well-instrumented workspaces — results depend on your categories, policies, and channels.

Up to 80%
Tier-1 deflection
depends on playbook + KB quality
Hours
Time to first value
channels + docs connected
Grounded
Knowledge-backed replies
citations from your sources
Gated
Destructive actions
human approval when you say so
Composite example

“We needed one queue, grounded answers, and approvals that didn't feel like IT tickets.”

Illustrative scenario — not a single customer endorsement.

The win wasn't “faster drafts.” It was fewer round-trips: the agent pulls order state, applies the policy we wrote, and only pings us when the amount crosses the threshold. Our queue stopped growing linearly with headcount.

−40–60%
queue pressure · first 60–90 days
Fewer tools
one inbox + automation layer
Faster replies
p50 improves when tier-1 is automated
Higher CSAT
when resolutions are consistent
Days
to first useful routing + KB
Talk to us about your rollout →
Changelog

Shipping iteratively. On purpose.

Updates on the blog →
Apr 24
shipped
Workflow publishing + snapshots
Immutable published definitions so production runs match what you reviewed.
Apr 18
shipped
Channel depth (email + chat + WhatsApp)
Bring conversations into one thread model with shared context.
Apr 11
shipped
Knowledge base ingestion
Chunk, embed, and cite sources so agents stay grounded.
Apr 03
preview
Insights + intent reporting
See what drives volume and where automation lands.
Pricing

Plans for agent seats with bundled automation.

Starter and Team include AI auto-resolves, channels, knowledge bases, and workflows. Enterprise adds the security and deployment options larger orgs expect.

Starter
From $24/seat/mo · billed annually
Small teams getting serious about support.
  • Shared inbox + AI agent
  • Bundled auto-resolves / month
  • Knowledge base
Compare plans
Most teams
Team
From $56/seat/mo · billed annually
Unlimited channels, SLA policies, deeper automation.
  • Unlimited AI agents (fair use)
  • Workflow automation + routing
  • Priority support options
Start free trial
Enterprise
Customannual contract
When security, residency, and procurement matter.
  • SSO / SCIM patterns
  • Custom terms + migration help
  • Solutions support
Talk to sales
Volume discounts at scale · Non-profit programs where available · 14-day trial on commercial tiers (see pricing page for details).
Why Subport exists

We care about the boring outcome: the ticket is resolved, the customer is unblocked, and the org can prove what happened.

Too many “AI help desks” stop at text generation. The hard part is safe action: tools, approvals, audit trails, and a queue model that doesn't fall apart when you add WhatsApp, email, and chat.

That's what we're building — one place for support to run, with automation that earns trust.

If you have a sharp edge case, bring it. We'd rather lose a deal than sell a story we can't ship.

Security & compliance

Boring on purpose.

Details live on the security page — here's the shape of how we think about risk.

Encryption
In transit + at rest
TLS for transport; strong defaults for stored data.
Access
Least privilege
Role and group patterns that match how teams actually work.
Auditability
Tool + config trails
Know what changed, what ran, and who approved it.
Data handling
Clear boundaries
Customer content stays scoped to tenant workflows.
Reviews
Ongoing
We treat security work as product work — not a one-time checkbox.
Enterprise
Procurement-friendly
Talk to us about DPA, SSO patterns, and deployment constraints.
Regions
Where you operate
Ask about data residency needs for your industry.
Transparency
Questions welcome
We'll walk security teams through architecture honestly.
Read the security overview →
FAQ

Straight answers.

A few things teams ask before they roll Subport out beyond a pilot.

Yes — start from the dashboard on a trial where available, or book a demo if you want a guided walkthrough.