Known issues from the knowledge base; outages to engineers

The question this file answersDo we really need an engineer on every ticket when most already have an answer in our knowledge base?

Fits: software and IT services companies handling 50–500+ tickets a day in Zendesk, Freshdesk, Intercom or a custom portal, with a knowledge base that already answers a good share of them.

Not for: teams with a few tickets a day, or products so new that most tickets are genuinely novel bugs — there is no known answer to point at yet.

Typical day

What the desk looks like today

Typical, from the software playbook — not a client's day. Tickets reach a mid-size SaaS or IT services firm at 50–500+ a day, through email, the portal and chat. An L1 engineer reads each one, reproduces it or searches the knowledge base, sets priority and type, drafts a reply and escalates what stumps them. Password resets and 'how do I export' sit in the same queue as a genuine outage. Release days and every new enterprise customer push the queue up, and so far the answer has been another L1 seat.

What changes

What Monday looks like after

07:45, before stand-up. The support queue opens with the repeat tickets already answered from the knowledge base and closed, or queued for a one-click send where a category is still under review. What is left for the engineer on rotation is the short list: a possible regression from yesterday's release, two tickets the classifier could not place, and an enterprise customer with their history at the top. Their day goes on those, not on the fortieth export question. You can see, per category, what closed without a person and what was reopened. Another support seat stops being the answer to every growth curve; the engineers still own every unknown.

Typical, not a measured client result. Every figure here comes from the playbook source named below.

~40–60%

faster resolution on routine tickets — Freshworks/Zendesk (2023)

Before: password resets and novel outages share one queue. After: with triage in front of the queue, Freshworks/Zendesk (2023) put L1 resolution 40–60% faster — engineers see the unknown, not the FAQ.

Where this number comes from

Freshworks/Zendesk benchmarks (2023) report 40–60% faster resolution on L1 issues with AI-powered triage; Gartner (2023) puts AI-handled IT support interactions without escalation at 30–40%. Vendor and analyst figures, not our result; software playbook range 45–65%. Not a CSAT or headcount claim.

What we install

What we put in front of the systems you already run

Customers keep reaching you as they do now — Zendesk, Freshdesk, Intercom or your portal stays the front door. In the queue we add a triage step through the tool's API:

each new ticket is classified and matched against the knowledge base
where a known answer exists, a reply with the KB link is drafted — sent automatically only for categories you have approved, otherwise queued for a one-click send
tickets that look like a new bug, an outage or an unhappy customer are routed to the right team with a history summary
escalations carry the reproduction steps
tickets the classifier is unsure about go to L1 as today.

First scope: the five repeat categories you close most.

What stays human — and what this will not do

Novel bugs. Complex troubleshooting. The relationship with the account. Feature requests that are really product decisions.

Not a CSAT or headcount claim.

What can go wrong — and what we do about it

A stale knowledge base means confident wrong answers at scale — the audit checks the KB against the last month's resolutions, and categories with weak articles stay manual until the articles are fixed. Custom portals without an API mean polling or a mailbox integration, slower and losing some fields. A customer who writes politely about a critical failure looks like a routine question; anything mentioning downtime, data or money goes to a person regardless of wording. The 45–65% range is for L1 and repeat issues; novel bugs, complex troubleshooting and feature requests are not in it.

How long it takes, and what we need from you

Audit, about two weeks (€1.5–3K): we read a month of tickets, count repeats with a KB answer against new issues, and check the knowledge base is worth pointing at. Pilot, 2–4 weeks for €10–20K, low-to-medium complexity by the software playbook: the top repeat categories, drafts reviewed by an engineer daily before any category goes to auto-send. Production: more categories, routing to teams, history summaries. From you: API access to the ticketing tool and the KB, and the customers who always get a person.

The path: free 60-second estimate → free 20-minute review → paid audit of this one process (€1.5–3K, typically two weeks) → pilot with your people in the loop (€10–20K, weeks, not quarters). No transformation programme. Prices are public, on the services page →

This is about you if…

Do tickets arrive by email, portal and chat into one queue with no sorting?
Do engineers or L1 staff read and answer every ticket, including the repeats?
Is there a knowledge base that already answers a good share of them?

What does this mean in euros?

That depends on your volumes and wage costs — this page will not invent the number. The free 60-second estimate runs that calculation from your answers, with every multiplier sourced.

Get your free savings estimate 60 seconds · no sales call Or write first → Map an ops process like this one — free, 60 seconds →

Not a named Aperanda client. Process file · Ops.

Deep-dive process file. Volumes, weeks and sources come from the industry playbook; nothing here is a named client.

All process files