Skip to content

Hjarni for Support Teams

Resolutions, product quirks, policies, and escalation paths in one place. So every agent's AI answers from the same source of truth, not from memory.

Free to start. No credit card required.

Product example

Answer the edge cases from your real policy

When your policies, product quirks, and escalation paths live in Hjarni, every agent's Claude or ChatGPT answers the in-between cases the way your team actually decided.

Claude
How do we handle a refund past the 30-day window?
hjarni: notes-get
Your refund policy note: past 30 days it needs a lead's approval. Store credit is the default unless the customer asks for the original method, and every exception gets logged in the Escalations folder.
Who approves the exceptions?
Any team lead. The policy note names Sam and Dana as the current approvers for EU accounts.

Claude answers a policy edge case using the support team's shared notes in Hjarni.

The problem

The easy tickets have macros. The hard ones do not. A refund past the window, a product quirk only two people understand, an edge case in the pricing policy. The answer lives in a senior agent's head or a thread from six months ago, so new agents guess, escalate, or answer inconsistently.

Put the real answers where every agent's AI can read them

Keep your resolutions, product quirks, policies, and escalation paths in a shared team space. Connect Claude or ChatGPT through MCP, and every agent's assistant answers from the same notes. Not a canned macro, the actual decision your team made.

What a support team keeps shared

  • Policies folder: refunds, exceptions, edge cases, and who can approve them
  • Product quirks: the known gotchas and workarounds that never made the help center
  • Resolutions: how the hard tickets actually got solved, kept for the next time
  • Escalation paths: who owns what, and when to hand a ticket up

Consistent answers, faster ramp

When every agent's assistant reads the same notes, "how do we handle a refund past the window?" gets the answer your policy actually says, every time. A new agent connects their own AI and ramps faster, because the team's hard-won context is one question away. Roles let a new agent read everything while only leads change policy.

The best answer lives in one agent's head. It doesn't have to.

What support teams keep in Hjarni

  • Policy and exceptions: the rules and the edge cases, with who can approve what
  • Product knowledge: quirks, workarounds, and known issues agents hit daily
  • Resolutions library: how hard tickets were solved, reusable next time
  • Escalation and ownership: who to reach, and when
  • Team AI instructions: folder-level rules every agent's assistant follows

Not a help center, not macros

Help centers are written for customers. Macros are canned text. Hjarni is the internal source of truth your agents' AI reads to handle the messy cases in between. Put it in a shared team space and every agent's Claude or ChatGPT answers from the same place. See Hjarni for Teams, or the set up your team guide.

Starter template

Skip the blank page

Paste a template link into Claude or ChatGPT and it builds the folders, tags, and AI instructions for you, so you start by adding notes instead of setting things up.

Common questions

Questions support teams actually ask

How is this different from a help center or macros?

Help centers are written for customers and macros are canned text. Hjarni is the internal source of truth your agents' AI reads to answer the messy, in-between cases: product quirks, edge-case policies, and escalation paths.

Does every agent need their own AI subscription?

Yes, and they bring the one they already use, like ChatGPT or Claude. Hjarni gives all of them the same shared knowledge base, so you are not paying per message and nobody is locked to one model.

Can we control who edits the knowledge base?

Yes. Viewer, editor, and admin roles, set team-wide or per folder, so a new agent can read everything while only leads change policy notes.

Will answers stay consistent across the team?

That is the point. When every agent's assistant reads the same notes, the answer to "how do we handle a refund past the window?" is the same one your policy actually says, not whatever each person remembers.

Start here

Write once. You both remember.

Free to start. No credit card required.

Works with Claude and ChatGPT today.