Issue Raised
Enrich with company dataCreate record
Notify owner
How can I improve support triage?
ArtifactUnderstand the issue, identify the priority, and route it to the right owner.
Enrich with company dataCreate recordNotify ownerLibra WorkBase
WorkBase for Customer Support
Every ticket comes with context, an answer, and a next step. WorkBase helps support teams resolve issues faster without making agents chase the work around the ticket.
Support agents are becoming human routing systems.
Tickets don't pile up because agents are slow. They stall because triage, research, follow-ups and next steps all happen after the conversation — usually while the queue keeps filling.
Where a support agent's 40-hour week goes
22.5 hours a week on non-customer conversations. That's 55% of a support agent's time spent moving work instead of resolving it.
The ticket lands. The work doesn't wait.
One trigger. The work is done. The reply is ready, the ticket is updated, and the next step is underway.
SlackMessages postedLibra
Refund request received
Customer is requesting a refund for a duplicate charge.
Order ID#A19288
PriorityHigh
Posted to #supportView message
GmailDraft readyTomansi@libra.com
SubjectRefund Request
Hi team, customer is requesting a refund for a duplicate charge. Order ID: #A19288.
Refund verified against the charge
Reply drafted for the customer
Ticket updated and routed
Drafted by LibraSend
Pick a workflow. WorkBase handles the busywork.
Choose a repeatable support task. Connect the systems it needs. Give it an instruction once. WorkBase keeps the workflow moving.
Ticket CreatedSummarise decisionsDraft the emailUpdate status
How can I draft better support replies?
ArtifactPull the relevant context and draft a clear response grounded in your support knowledge.
Summarise decisionsDraft the emailUpdate statusCustomer escalationPull ticket statusRead recent threads
How can I prepare an escalation review?
ArtifactPackage the issue, customer history, reproduction details, and impact before handing it over.
Pull ticket statusRead recent threadsFlag at-risk
Your CRM stores the deal. WorkBase runs it.
Salesforce, HubSpot and Pipedrive are systems of record. They tell you what happened. WorkBase reads from them and knows what to do next.
| Dimension | Traditional CRM | Libra WorkBase |
|---|---|---|
| Ticket triage | Agent reads and routes | Auto-updated from email, meetings and calls |
| Follow-ups | Agent researches and writes | Drafted from the meeting context, ready to send |
| Escalations | Explain the issue again | Context travels with the handoff |
| Customer history | Spread across systems | Connected around the ticket |
| Follow-through | Agent remembers | Workflow keeps moving |
Give every ticket the whole story.
WorkBase pulls together the customer, conversation, product, and support context an agent needs without making them search for it.
Customer history
Knowledge-based answers
Engineering context
Auditability
Connectors for support
Salesforce
Hubspot
LinearGmailSlack
Google Drive
Notion
Zoom
Before you connect anything.
No. Salesforce, HubSpot and Pipedrive stay your system of record. WorkBase reads from them and writes back to them, so agents stop typing what the tools could have captured.
A plugin fetches a record when you ask. WorkBase runs the whole task: it reads the ticket, the thread and the CRM together, drafts the reply, updates the case, and sets the next step — under your rules, not one lookup at a time.
Only where you allow it. You decide which fields and actions run automatically and which wait for a one-click approval, and every write is logged so you can see exactly what changed and why.
Connect your helpdesk and CRM through scoped OAuth, pick a task, and give it one instruction. It works on day one — no dashboards to build and no Zapier chains to maintain.
Yes. On Enterprise the full stack can run inside your own environment, so indexed content and connected data never leave your walls.
No. Workspace controls scope what each person sees: support leads see everything, an agent sees only their own queue, and connected sources keep their own permissions.