Job helpdesk-tag-list-cleaned-and-defined · revised 12 October 2026
Cut your help-desk tags to an agreed list, with a definition and two examples each
Your tag list cut to an agreed size, each tag with a one-line definition and two examples, an old-to-new mapping and a retire list, checked by two of your staff on 50 tickets.
You might be seeing
- Dozens of tags have been used once or never
- Near-identical tags split the same topic
- New staff cannot say which tag to use
No passwords, keys, card details or admin invites needed to start.
What usually happened
A tag list grows by accident until staff cannot tell which tag to use, so the same issue is tagged three ways. Counts by tag then say little, and any later automation inherits the confusion.
Who it’s for: Owner or support lead whose help desk has hundreds of tags that nobody can explain.
Usually starts when: Two people tag the same ticket differently, and reports by tag mean different things in different months.
The result: Your tag list is cut to an agreed size and each remaining tag has a one-line definition and two examples. You also receive an old-to-new mapping, a retire list, and a check that two of your staff tag the same 50 tickets with the new list and agree.
Check whether this job fits
Answer these without sending chats, tickets, files or logins. Nothing is submitted unless you choose to contact us.
What you get
- Usage counts by tag
- The proposed list with definitions and examples
- The old-to-new mapping and retire list
- The 50-ticket agreement result
Included
- One help desk and its tag export with usage counts by tag
- Merge, keep and retire proposals, each with the count behind it
- A one-line definition and two real, redacted examples for each kept tag
- An old-to-new mapping for every retired tag
- A 25-ticket trial and a separate 50-ticket check by two of your staff on tickets we have not seen, with agreement reported
Not included
- Applying the new list or re-tagging past tickets in your help desk
- Deciding what to retire: your decider does that
- Automatic tagging of new tickets: see the ticket suggestion job
- Changes to your help desk by us
- Voice or call tags
How we know it’s done
Agreed with you before work starts. Each check produces evidence you keep.
Two of your staff tag the same 50 tickets (chosen by your decider, unseen by us until the final count, and none used as an example in the tag sheet) with the new list and agree on at least 85% of them (raw agreement and the confusion table are shown, not kappa alone; 43 of 50 shows a true agreement rate of at least 75% with 95% confidence).
Evidence: The 50-ticket table with both staff members' tags.
Every retired tag has a mapping to a kept tag, and counts reconcile to the export (old counts of merged tags add up to the new tag's count).
Evidence: The mapping table and a reconciliation of counts to your export.
Every kept tag has a one-line definition and two examples that your decider has read.
Evidence: The tag sheet with your decider's approval.
Sign-off. Your decider reads the list, the mapping and the agreement result and signs off in writing. Payment follows sign-off.
If it fails. If staff cannot reach the agreement mark after one rewrite of the definitions and a count on 50 fresh tickets, we hand over the list and the confusion table and you decide whether to stop or re-quote.
When it fits, and when we stop
It fits when
- A tag export with usage counts can be produced from your help desk
- A named person decides which tags merge or retire
- Two of your staff can tag 25 trial tickets and then 50 more independently with the new list, and 50 further fresh tickets if the first count fails
We stop and tell you if
- No usage data exists for the tags
- No one can decide which tags to merge
- The two staff disagree on more than 15% of a first 25 tickets: we rewrite the definitions and you pick 50 fresh tickets, unseen by us until the final count
What could go wrong
We only read an export and write a list, so nothing in your help desk changes. Applying the list is your decision and the old tags can be kept until you are ready.
Scroll the table sideways to read it all.
| Risk | How we handle it |
|---|---|
| A merge hides a distinction that matters for reports | Each merge is shown with its counts and approved by your decider. Retired tags keep a mapping so nothing is lost. |
| Staff still disagree after the clean-up | Raw agreement and the confusion table on 50 tickets are reported, and definitions are rewritten where two staff differ. |
| Real ticket text reaches us | Examples are redacted by you before sharing, through a secure route agreed in writing. |
A separate automated reviewer checks the counts, the mapping and the 50-ticket result before hand-over, and your decider on your side marks the sample and decides what goes live. No human supervisor is included unless your agreement names one. At launch the work is largely automated, and we say so.
How we deliver
We arrange the work and independent review, then show you the result against the agreed checks. You keep authority over your systems.
- Agree the help desk, the export and who decides, in writing
- Count tag usage and group near-identical tags with the evidence
- Write a definition and two redacted examples for each kept tag
- Your decider approves merges and retirements; we build the mapping
- Your two staff tag 25 trial tickets, then 50 fresh ones; we report agreement
- Independent automated review; hand over the list and mapping
This is a one-off job, not emergency cover or a subscription. We confirm eligibility, the total price, a start window and a delivery date before you accept. Work starts only after agreed inputs, secure access and any needed permissions are in place. Model usage for our build and tests is included in the price; charges on your own platform account, including per-conversation fees after go-live, are yours. An enquiry creates no charge or booking.
Need to keep it working?
Discuss suggestions for new tickets using the cleaned list.
Ongoing work is separately scoped and quoted: no monitoring, response-time guarantee or automatic subscription is included in this job.
Explore an ongoing managed lane, or mention the responsibility you need in your enquiry.
What you can check
This is a new service. We have not delivered this job for a client yet.
Other ways to get this done
- Export the tag counts, sort by use, and merge any tag used fewer than five times into its nearest neighbour. That is a free first pass.
Questions
Do you change my help desk?
No. You receive the list and the mapping, and you apply them.
Is any AI used?
The analysis may use a language model to group tags and draft definitions. Nothing runs inside your help desk afterwards.
What comes next?
A clean list is the first step before automatic suggestions for new tickets.
Send an enquiry
Send us
- Your help desk and about how many tags it has
- Whether anything reports by tag today
- Who decides what a tag means
- No customer data, logins or keys in the first enquiry
Later, once you agree
- A tag export with counts, and a redacted sample of tickets for the examples
- Two staff to tag the same tickets: 25 for a first trial, then 50 different ones chosen by your decider and unseen by us until the final count
- A secure route agreed in writing before any ticket text is shared
You keep the help desk and every login. We work from an export you create with personal details removed, and we change nothing in your help desk. You apply the new list yourself.
Email fallback: open your mail app
If website submission is unavailable, review and send the fallback email yourself. An email fallback is not a website receipt. Or write to hello@syntheticindustry.ai with “helpdesk-tag-list-cleaned-and-defined” as the subject.