New · SupportCross-functional
Tier-1 tickets, resolved. Or escalated with the full story.
Reads every incoming ticket against your help center, product docs, and years of past resolutions, then closes it with a cited answer or hands it to a human with the research already done — never a dropped thread, never a customer re-explaining themselves. Built for support teams drowning in repeat questions.
Every resolution cited to a source article or prior ticket. Built on Sphere AI Foundry.
support_triage — ticket #48213LIVE
00:00Ingest → "Can't reset 2FA after phone upgrade"read
00:02Search KB → 3 candidate articles, 1 similar past ticketmatched
00:04Confidence check → 0.94, above 0.85 auto-resolve thresholdpassed
00:06Draft reply → backup-code recovery steps, citedpassed
00:07Outcome → ticket closed, customer notifiedresolved
resolution time 7stickets auto-closed 61%human review on escalation
Built on Sphere AI Foundry·Zendesk, Salesforce & Freshdesk ready·21 years, 300+ clients
Manual review doesn't scale with volume
Most tier-1 tickets aren't new problems — they're the same password reset, billing question, or setup issue your team has answered a thousand times before, buried in a queue alongside the handful that actually need a person. Agents spend their day searching the same help center and re-typing the same answers instead of handling the tickets that need judgment.
61%
of tier-1 tickets auto-resolved without a human touch
7s
median time to a cited first response
100%
of resolutions traceable to a source article or ticket
Ticket in, cited resolution out
Four steps, every one logged and citable back to a help center article or a past ticket.
01
Ingest
Reads the ticket directly from your queue
New tickets, replies, and reopens are read straight from Zendesk, Salesforce Service Cloud, or Freshdesk as they arrive — no separate triage inbox, no manual tagging step.
Source ticket
#48213 — "Can't reset 2FA after phone upgrade"
02
Search knowledge & history
Checked against your docs and every past resolution
Help center articles, product documentation, release notes, and the full archive of previously resolved tickets are searched together — not a static FAQ, the actual institutional memory of your support org.
Example query
"How does a customer recover 2FA access after losing their old device?"
03
Cite every answer
No resolution without a citation
Each answer links back to the specific help article, doc section, or prior ticket it was drawn from — the citation ships with the reply, so a reviewer or the customer can verify it directly.
Cited resolution
Backup-code recovery is the fastest path; escalate to identity verification only if none exist.
cited to KB-1142, Account Recovery: 2FA04
Resolve or escalate
Closes the ticket, or hands it off with the research done
Above your confidence threshold, the agent replies and closes the loop. Below it, or against an escalation rule, it routes to a human agent with the ticket, the candidate answers, and its reasoning already attached.
Outcome
Auto-resolved ✓Escalated, context attachedNo match found
10 fields, every one cited
What the agent produces for every ticket it touches.
Intent & category classification
Confidence score
Cited draft reply
Source article or ticket reference
Customer & account context summary
Prior-ticket history on the account
Escalation reason, when routed
Suggested tags & priority
Knowledge-gap flag for missing docs
Resolution audit log
Typical inputs
What the agent reads, and what it connects to.
Documents
- Help center and knowledge base articles
- Product documentation and release notes
- Resolved ticket history and macros
- Escalation and routing rules
- Service-level agreements by customer tier
Systems
- Zendesk, Salesforce Service Cloud, or Freshdesk
- Help center / knowledge base platform
- CRM for account and entitlement lookup
- Slack or Teams for escalation handoff
Is this the right fit?
Built for volume — not a replacement for judgment calls.
✓ Works best for
- Support teams fielding hundreds of repetitive tier-1 tickets a week
- Organizations with an established help center and ticket history to learn from
- Teams wanting every auto-resolution traceable to a source
Too small for
- A support desk fielding a handful of tickets a week a single agent easily keeps up with
- Replacing human judgment on billing disputes, churn risk, or angry customers
- Teams without a knowledge base or documented resolution history to ground answers in
Grounded in your knowledge base, not a black box
Compliance-first, the same standard every Sphere agent is held to.
The agent only auto-resolves a ticket when it finds a documented answer above your configured confidence threshold — anything ambiguous, anything matching a sensitive category like billing disputes or account cancellation, and anything it simply can't ground in a source routes straight to a human agent, never a guess sent to a customer.
Every escalation carries the full research trail with it, so no context is lost on handoff — see how the same governed, cited-answer standard applies across regulated deployments in Governed AI for Banks and Private LLM Deployment for Banks and Financial Institutions.
Prerequisites knowledge base
A maintained help center and a history of resolved tickets to ground answers in before go-live.
Human review on escalation
Anything below the confidence threshold or matching an escalation rule routes to a person with full context.
Regulatory grounding policy-aware
Escalation rules encode your sensitive-category policies — billing, cancellation, legal — directly.
Foundation Sphere AI Foundry
Access controls and audit trail configured here carry forward to every other agent you deploy.
Frequently asked
Does the Support Knowledge Agent replace tier-1 support agents?
No. It closes the tickets that have a clear, documented answer and hands off everything else — ambiguous requests, angry customers, account-specific exceptions — to a human agent with full context attached. It's built to absorb repetitive volume, not to run your queue.
What ticketing platforms does it work with?
Zendesk, Salesforce Service Cloud, and Freshdesk out of the box, reading directly from your existing queue. Other ticketing platforms can be added during onboarding.
How does it decide when to answer versus escalate?
It only closes a ticket when it finds a documented answer with a confidence score above your configured threshold, cited to a specific help article or a prior resolved ticket. Anything below that threshold, or anything matching your escalation rules, routes to a human with the full research trail attached.
Is every answer traceable to a source article or ticket?
Yes. Every resolution cites the specific help center article, product doc, or prior ticket it drew from — there is no answer sent to a customer without a citation behind it.
How is this different from a generic chatbot or help widget?
A generic chatbot answers from a static FAQ and stalls outside it. This agent is built around your actual ticket history, product docs, and escalation rules, and it runs on Sphere AI Foundry, so the access controls and audit trail configured here carry forward to other agents you deploy.
Start here
See it on your own ticket queue
Bring a week of real tickets to the demo. We'll show the resolution output live — classification, citations, and the escalation handoff, on your own queue.
Looking for a different workflow? Browse the full Agent Catalog.
Talk to a solutions architect
Direct to a senior architect — never a sales queue. Replies within one business day.
By submitting, you agree to be contacted about this request. We never share your details.