L2 support

L2 support is the second tier in a support escalation model: agents who handle cases that need more product knowledge, system access, or judgment than L1 is equipped to resolve.

A case reaches L2 when it does not fit an L1 script, but it also does not need engineering or specialist policy input, which would push it to L3 instead. In financial services, L2 often means the agent has access to more back-office systems and more discretion, such as authority to approve a goodwill payment or open a full dispute investigation.

When does a case escalate from L1 to L2 support?

Escalation from L1 to L2 typically happens on three signals: the query falls outside documented L1 procedures, the case needs system access or authority L1 does not have, or the customer has already tried the standard steps and the issue persists. An AI agent scoped to L1 should recognise these signals and hand off rather than attempt a workaround, because acting outside its documented scope is where errors and compliance risk concentrate. A well-built escalation path makes this handoff visible in the case record, not silent.

Frequently asked questions

What access does L2 typically have that L1 doesn't?

L2 usually has broader system access, higher approval limits, and the authority to make judgment calls L1 cannot, such as approving an exception or opening a formal investigation. The exact scope varies by firm and case type.

Does every customer query pass through L1 before reaching L2?

Not always. Some queries are routed directly to L2 based on the contact reason, such as a dispute filed through a dedicated channel, without needing an L1 triage step first.

Is L2 support more expensive to run than L1?

Generally yes, per case, because it needs more experienced staff or a more capable AI agent with wider system access. This is why routing accuracy between L1 and L2 matters: sending a simple case to L2 unnecessarily costs more than it needs to.