Standing service free-text-tagging-checked-against-your-labels-every-month · revised 12 October 2026
Standing service
Every month, tag new comments or tickets in my categories and show how well it agreed
Each month new rows of free text are tagged with your labels, and 30 rows you label are used to report agreement and any change since last month. Stored decisions are reused.
This starts a conversation by email. Nothing is charged, and nothing is run, until we have agreed scope, data route and terms with you in writing.
The responsibility you hand over
A tagging step that worked in month one can drift as the wording of comments changes, a label shifts in meaning or a provider updates its model. Without a fresh check against your own labels each month nobody sees it until a report looks wrong.
Who it’s for: Operations or customer-experience analyst who tags a new batch of comments or tickets every month.
Usually starts when: Each month's batch is tagged by hand, or tagged by a tool nobody has checked since the first month.
The result: Every month your new rows are tagged with your labels, and 30 rows you label are scored for agreement. The report shows agreement, per-label counts and any change since last month. Stored decisions are reused, so old tags do not change.
What stays true, and what we do about it
No response-time guarantee is published for this new service. A date for each month's report is agreed in writing before it starts, set to what a service at this stage can keep.
Runs once a month after the sample arrives. At launch the service is not staffed round the clock.
What must remain true
- Each month's rows are tagged only with labels from your list or 'unsure'
- Agreement with your labels on 30 fresh rows is reported, with any fall since last month
What we watch
- The month's redacted file
- Your labels for 30 fresh rows
- The label guide
When something happens
Scroll the table sideways to read it all.
| When | What we do |
|---|---|
| A month closes and the file and your 30 labels arrive. | We tag the file, score the 30 rows and send the tagged file and the report. |
| Agreement falls below the floor you set. | We list the labels and rows that changed, and propose a guide change or a re-tune. You decide. Linked job: Free-text rows tagged with your labels |
| You change or add a label. | We update the guide and re-score the sample before using the new label. |
We do on our own
- Tag the file with the frozen script
- Score the 30 rows
- Prepare the report
We ask you first
- Any change to a label definition or the floor
- Retuning the script
- Any request for wider access, for a login or for your provider key
We escalate to you when
- Agreement falls below the floor two months running
- A label with at least 10 sampled rows over the last three months has recall below 60%
- The file's shape or language changes
How you know it held. Each month you can recompute agreement from the 30 rows and the tagged file, and the report lists the tags that changed since last month.
How we keep it true
This service is never finished. Each month's report shows whether the thing it watches stayed true, and it continues until you end it.
Sort my comments or tickets into my categories, and show the tags stay stable on a re-run Job Optional
Each row of comments or tickets gets one label from your list, or 'unsure'. Agreement with your labels and stability across runs are measured. Short descriptions: see the text-column job.
Bought separately; not part of the monthly fee
The one-off job that builds the label guide, the frozen script and the first agreement test.
What is included, and what is not
- The month's tagged file
- The agreement report
- The unsure list
- The updated label guide, if you change a label
- The list of tags that changed, if any
Included
- One monthly file of up to 10,000 rows of text and your list of up to 20 labels
- Tagging with the frozen script from our tagging job or your own
- 30 fresh rows a month that you label, for the agreement check
- A monthly report: agreement, per-label counts, unsure share and changes
- Updates to the label guide when you change a label
Not included
- Labels that judge people, or legal conclusions
- Personal data in the text
- Hiding or removing customer reviews
- Out-of-hours cover or any guaranteed response time
- A promise that agreement stays at today's level
How we know it’s done
Agreed with you before work starts. Each check produces evidence you keep.
Each month the report shows agreement on 30 rows you labelled, per-label counts (not percentages, because 30 rows are too few for them), and the unsure share.
Evidence: The report and your 30 labels.
No label outside your list appears in the tagged file.
Evidence: A script check.
Previously stored tags are unchanged unless you approved a change.
Evidence: A before and after comparison.
Sign-off. You read each report and tell us whether the marks and rules still suit you. A month counts as delivered when you have received the report.
If it fails. If we cannot run a month because a sample is missing or has changed shape, we say so, explain what we need and do not guess. If the service is not working for you, you can end it at the end of any month.
When it fits, and when we stop
It fits when
- A frozen tagging script and a label guide exist, from our one-off job or your own
- A named person labels 30 fresh rows each month, about 10 minutes in all
- The text holds no personal data, or is redacted at source
- Two people agree on the label definitions
We stop and tell you if
- The labels judge people
- Personal data cannot be removed from the text
- Nobody can label the monthly sample
- You want unsure rows forced into a label
What could go wrong
The service reads samples and writes reports, so there is nothing to undo. Ending it leaves your systems exactly as they are; you keep every report we sent.
Scroll the table sideways to read it all.
| Risk | How we handle it |
|---|---|
| Agreement drifts and nobody notices. | The 30-row check runs every month, and a fall below the floor is raised. |
| Instructions hidden in the text change a tag. | Text in rows is treated as data and tested at the start with rows you plant. |
| Your files go to a model provider outside your control. | Before any file is shared we name the provider, its country and its retention terms in writing. Entry jobs run on redacted, synthetic or public copies, or on your own provider account with a key you hold. |
A separate automated reviewer checks each month's agreement report and the changed tags before hand-over and does not mark the monthly sample. A named person on your side labels the 30 rows and signs off. No person on our side is included unless your agreement names one, and at launch the work is largely automated.
Stays with a person
- You or your named person accept every report and every change to a rule, mark or setting
- You apply anything the report suggests in your own systems
Access we would need
- Redacted, synthetic or public samples you send each month by an agreed secure route
- Your labels for the monthly sample
Questions
Will old tags change?
No. Stored decisions are reused unless you approve a change.
Why 30 rows each month?
They are a cheap tripwire for drift. They catch big changes, not small ones.
Do I need the one-off tagging job first?
Not if you have a label guide and a script. We check them in the first month.
Also part of
Send an enquiry
Send us
- Your label list and how many rows arrive each month
- Whether a frozen tagging script exists
- No real records, personal data or credentials in the first enquiry
Later, once you agree
- Each month's redacted file by an agreed secure route
- Your labels for 30 rows
- A named person who decides the unsure rows
- A written cap on model usage
You keep your files and every decision. We work on the redacted file you send each month, return the tagged file and report, and delete our copies at the end of each month. We take no logins to your helpdesk or survey tool.
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 “free-text-tagging-checked-against-your-labels-every-month” as the subject.