Help desk, ticketing, live chat and knowledge base, with AI agents that answer, investigate and take action. Switch from Zendesk or Intercom without a cutover.
#7265·Different SLAs per customer tier?
ResolvedRWe have customers on our Pro plan who matter way more to us than others, and we want tighter SLAs for them. Can we set SLAs per customer tier? Our top 20 accounts get 1 hour, everyone else gets 4?
Hi Ben. Per-customer SLA overrides aren’t available today: SLA policies attach to plans, so every account on a plan gets that plan’s SLA. There’s a workaround that gets you most of the way. Want me to walk through it?
Yes please
It uses conversation types:
- Create a conversation type called “Priority Account.”
- Add a routing rule: if the company is tagged as a top account, set the type to Priority Account.
- Attach a 1 hour SLA to that type. Type SLAs take precedence over plan SLAs.
All 20 share one SLA, so three tiers would need three types and three rules.
That works for now. Let’s use a tag called “tier-1”. I’ll add it to the companies myself today.
StripeHalcyon Media · Pro plan, renews Jan 14
HubSpotOwner: Tom · 20 accounts tagged “top account”Hey Tom, I found a few things you might want to look at.
Ben wanted per-customer SLA targets, which don’t exist. The Support Agent said so plainly, built the conversation-type workaround once he picked a tag, and logged the request.
- Tagging a company
tier-1now gives it a 1 hour response target - All tier-1 accounts share one policy, not a target each
- The
7th requestin 90 days, across5 accountsand$214k ARR
Your Feature Request Agent found something worth a look.
Halcyon Media Per-customer SLA overrides. The 7th requestacross5 accountsin 90 days,$214kof combined ARR.- Matched to 6 existing requests
Your Revenue Agent found something worth a look.
- Halcyon Media is on
Proand asking for per-account SLA targets, an Enterprise capability. - They’ve already named their
top 20 accounts, so the requirement is scoped.
Your Knowledge Agent found something worth a look.
- No help article covers
tighter SLAs for specific customers. - The same workaround was typed out by hand in 7 tickets across 5 accounts.
That’s live now, Ben. Tag a company “tier-1” and anything from them picks up the 1 hour response target.
One limitation worth repeating: all tier-1 accounts share the same policy. Three different targets would mean three conversation types and three rules.
Your request is on file with six others asking for the same thing. I’ll come back to you if it moves.
Trusted by 1,400+ companies globally
Your support queue already knows who's about to churn.
Background agents read every conversation and flag churn risk, expansion opportunities and feature requests, with the revenue at stake and a suggested next step. Support stops being a cost center and starts protecting revenue.
Churn agent
Spots frustration, downgrade talk and silent accounts before they cancel.
See how it worksFeature request agent
Groups every request by account and revenue so product knows what matters.
See how it worksThe problem
Three jobs stand between your team and the work that matters.
Helply takes all three off your team’s plate, so the conversations that need a person finally get one. We handle the busywork.
Resolved end to end
It closes about two-thirds of our tickets on its own, lookups and all.
David Blacker
Co-founder & CEO, AirGigs
Resolved with no drop in accuracy
It hit an ambiguous question and escalated instead of bluffing. That convinced us.
Razia Aliani
Senior Systematic Reviewer, Covidence
Fewer tickets after leaving Zendesk
I haven’t had a single panic moment since we switched.
Jacqueline Antworth
Director of CX, Proposify
Answering the same questions
Password resets, plan limits, order status
Chasing answers across tools
Stripe, the admin panel, the engineer in Slack
Following up after the ticket
Bug reports, quiet accounts, stale help articles
The conversations that need a person
Upset customers, edge cases, the relationships you built
Built for how support actually runs.
A full helpdesk with agents built in, not bolted on. They answer customers, draft for your team, and keep watch in the background.
Tickets that close themselves.
Helply answers from your docs and past resolutions, looks things up in your tools, and takes the action: the refund, the reset, the plan change. It escalates the moment it isn’t sure.
- Answers from what you knowDocs, macros and past resolutions, cited in every reply.
- Takes the actionRefunds, resets and plan changes, done through your tools.
- Knows when to hand offEscalates to a person the moment it isn’t sure.
Every reply arrives as a finished draft.
For the tickets that need a person, the investigation is already done. Your team reviews, edits if needed, and sends.
- The investigation is doneLogs, account data and similar tickets, gathered before anyone opens the ticket.
- Drafts in your voiceReview it, edit if needed, and send.
- Summaries on demandLong threads boiled down in one click, with sources.
The helpdesk your team already knows.
Shared inbox, SLAs, views, assignments and reporting. Everything your team expects, with AI working inside it.
- One shared queueEmail, chat, Slack Connect and phone in one place.
- SLAs and routingTimers, assignments and views your team already knows.
- ReportingVolume, resolution time and AI performance side by side.
Support agent: a seat refund checked, issued and resolved in 23 seconds.Assist agent: the investigation done and a reply drafted before anyone opens the ticket.Inbox: every ticket, SLA and AI draft in one shared view.
Helply answers from your docs and past resolutions, looks things up in your tools, and takes the action: the refund, the reset, the plan change. It escalates the moment it isn’t sure.
- Answers from what you knowDocs, macros and past resolutions, cited in every reply.
- Takes the actionRefunds, resets and plan changes, done through your tools.
- Knows when to hand offEscalates to a person the moment it isn’t sure.
Support agent: a seat refund checked, issued and resolved in 23 seconds.
For the tickets that need a person, the investigation is already done. Your team reviews, edits if needed, and sends.
- The investigation is doneLogs, account data and similar tickets, gathered before anyone opens the ticket.
- Drafts in your voiceReview it, edit if needed, and send.
- Summaries on demandLong threads boiled down in one click, with sources.
Assist agent: the investigation done and a reply drafted before anyone opens the ticket.
Shared inbox, SLAs, views, assignments and reporting. Everything your team expects, with AI working inside it.
- One shared queueEmail, chat, Slack Connect and phone in one place.
- SLAs and routingTimers, assignments and views your team already knows.
- ReportingVolume, resolution time and AI performance side by side.
Inbox: every ticket, SLA and AI draft in one shared view.
Support for every kind of customer.
Helply learns how your business works, then handles the questions your customers actually ask.
B2B SaaS and AI-Native
For teams handling hundreds to thousands of tickets a month who want AI that can look up accounts, check billing and fix things, not just deflect.
Marketplaces
Two-sided support with refunds, payouts and disputes. Agents take the action through your API or MCP.
Also used by consumer apps and fintech teams.
An AI agent beside every teammate.
The Assist agent works inside every ticket with your team. Ask it about the customer, the thread or your docs, and it answers with sources.
- Summarizes long threads in one click
- Drafts replies in your voice, with sources you can check
- Suggests the next step: refund, escalate, or loop in an engineer
Starts as a draft, then earns autopilot.
You stay in control.
Helply doesn’t take over on day one. It proves itself topic by topic, and you decide what it handles alone.
It learns your business
Connect your helpdesk and docs. Helply reads your past tickets, saved replies and help articles overnight.
You approve, it sends
Every reply starts as a draft your team approves in one click. You see exactly how it reasons before it ever goes live.
It earns autopilot
Turn on autonomous resolution one topic at a time, with your rules for when to escalate. Anything it isn’t sure about goes to a person.
One helpdesk. No bolt-ons.
Everything a support team needs, in one place and on one bill.
Shared inbox
Assignments, views, SLAs and collision detection.
Chat widget
Live chat on your site and in your app.
Knowledge base
A help center the agents keep up to date.
Every support address in one place.
Slack and Teams
Support customers where they already talk to you.
Phone and SMS
Calls and texts in the same timeline.
Workflows
Routing, tagging and automations without code.
Analytics
Resolution, CSAT and AI performance in one view.
Unlimited view-only seats from Team up. AI credits pooled across your workspace.
See pricingOther helpdesks bolt AI on. Helply was rebuilt around it.
| Legacy helpdesks | Helply | |
|---|---|---|
| AI | Sold as an add-on, priced on top | Built into every part of the helpdesk |
| Pricing | Per seat, so every new teammate costs more | Plans include seats and AI credits, with unlimited view-only seats from Team up |
| What the AI can see | The ticket in front of it | Your docs, past tickets and connected tools |
| What the AI can do | Suggest an answer | Resolve: refunds, resets, plan changes, escalations |
| Switching | Months of migration | Overnight import, history included |
Use your AI. Keep Helply as the system of record.
More teams are building support workflows in Claude. Helply's MCP server gives your AI the same access our agents have: tickets, customers, history and your knowledge base. Your AI reads, drafts and acts. Helply stays the place every conversation, SLA and report lives.
Claudevia Helply MCPConnect Claude in minutes
Point Claude or any MCP client at your Helply workspace. No custom integration to build or maintain.
Read and write, not just read
Search tickets, pull account history, draft replies, update statuses and tag conversations from your own AI.
Mix and match
Run Helply's agents where they're strongest and your own where you've already built something good. Same queue, same reporting.
Switch without a cutover.
Connect your support inbox and run Helply next to your current helpdesk. Keep your old tool for history while Helply handles new tickets. We import your help center, macros and saved replies during the trial.
Imported to Helply- Customers and companies
- Macros and saved replies
- Help center articles
- Tags, views and SLAs
Works with the tools you already run on.
The agents don’t just read your tools. They act in them: refunds in Stripe, bugs in Linear, deals in HubSpot.
SlackAnswers in Slack Connect channels
HubSpotContacts and deals in every ticket
StripeRefunds and billing lookups
LinearBugs filed from conversations
JiraEscalations to engineering
AttioCRM context for every customer
GitHubIssues linked to tickets
NotionDocs the agents can read
SalesforceAccounts and opportunities in every ticket
ZapierConnect 6,000+ more apps1,400+ teams run support on Helply.
“What convinced us was watching it decline to answer. It hit a genuinely ambiguous question and escalated instead of bluffing.”
“I haven’t had a single panic moment since we switched. The routine stuff resolves itself correctly in the background.”
“Support used to be a cost line. Helply turned it into our most reliable signal, every onboarding question routes back to product the same week.”
I’ve been building helpdesk software since 2011. Two years ago I realized AI wasn’t a feature we could add. It meant starting over.
So we did. We spent 18 months rebuilding Groove from scratch as Helply, a helpdesk where agents do the work and people handle the judgment calls. Then we ran our own support on it.
Our live queue went from six people to two, and the other four moved into sales and success. Nobody was let go. The work just got better.
That’s what I want for your team. Not fewer people. Better work for the people you have.
What is Helply?
Helply is an AI-native customer support platform that combines a complete helpdesk with AI agents. Your team manages customer conversations while AI agents investigate issues, resolve requests, take action, and work alongside your support team.
How is Helply different from a traditional helpdesk?
Traditional helpdesks were built around humans managing queues of tickets. Helply was built from the ground up for humans and AI agents to work together. AI agents can investigate issues, resolve conversations, update other systems, and handle repetitive support work while your team stays in control. It includes everything you expect from a helpdesk: shared inbox, messaging, knowledge base, analytics, customer records, and workflows.
Can Helply replace Zendesk, Intercom, Help Scout, or Freshdesk?
Yes. Helply is a complete customer support platform, not an AI add-on. Teams can manage customer conversations, tickets, knowledge, workflows, reporting, integrations, and AI agents in one platform. Compare Helply to Zendesk, Intercom, Help Scout, or Freshdesk.
What are AI customer support agents?
AI customer support agents are software agents that perform support work alongside your team. Helply's AI agents can understand customer requests, investigate problems, answer questions, take actions, update systems, escalate issues, and resolve conversations when they have enough confidence.
What is agentic AI customer support?
Agentic AI customer support uses AI agents that investigate a customer's issue, decide what to do, and take action. Helply is built this way: its agents read the conversation, check account and product data, run approved actions such as refunds, and hand off to your team when a decision needs a person. Every action runs inside permissions you set and is logged.
Can Helply’s AI agents take actions, or do they only answer questions?
Helply's AI agents can take actions. They can use connected tools, APIs, customer data, account information, product systems, and internal workflows to investigate and complete support tasks, not just generate answers. See AI agents and integrations.
Does Helply work for both B2B and B2C customer support?
Yes. Helply is built for both B2B and B2C customer support teams. AI agents can work across different support volumes, channels, workflows, customer types, and levels of complexity while giving human agents the context they need.
What customer support channels does Helply support?
Helply brings customer conversations into one support platform across channels including email, live chat, Slack, Microsoft Teams, SMS, WhatsApp, and other connected integrations.
How does Helply pricing work?
Starter is $79 a seat a month for up to 3 seats, with 500 AI credits per seat. Team, Pro and Growth cost $499, $999 and $1,999 a month and include 15,000, 50,000 and 100,000 AI credits. Credits are priced by effort: a drafted reply is 2, an investigation across tools 6, and an end-to-end resolution 10. See pricing.
Do we pay per user?
Yes, per seat. Starter is $79 a seat a month, and Team, Pro and Growth include 5 to 10 seats, then charge from $99 a seat. From Team up, view-only seats are unlimited, so teammates who only read and leave notes cost nothing.
Can I migrate my existing helpdesk to Helply?
Yes. Teams can migrate from customer support platforms like Zendesk, Intercom, Help Scout, Freshdesk, and Groove to Helply while bringing over the knowledge and customer context their team and AI agents need. Book a demo to plan your migration.
Is Helply an AI chatbot?
No. Helply is a complete AI-native customer support platform. Customer-facing AI is one part of the product. Helply also provides a full helpdesk and AI agents that can investigate issues, assist human agents, take actions, analyze customer signals, and perform support work behind the scenes.
The complete system for human and AI customer service.
Start with a 14-day free trial. No credit card, and a person from our team to help you set up.