Job crm-fake-form-leads-held-for-review · revised 12 October 2026
Hold bot leads from your web forms for review, without deleting any of them
Submissions that match junk rules you approve are held in a review view, not routed or counted. A test of 30 junk and 30 genuine invented submissions checks that no genuine one is lost.
You might be seeing
- Reps work leads with random names and addresses
- Lead counts jump with no matching rise in real enquiries
No passwords, keys, card details or admin invites needed to start.
What usually happened
Bots and spammers fill web forms, and the CRM treats every submission as a lead. Junk is routed to owners, enrolled in sequences and counted in reports. Blocking at the form can reject real people, so the choice feels like losing leads either way.
Who it’s for: Marketing manager or agency whose paid campaigns bring form submissions that are mostly junk.
Usually starts when: Junk submissions are routed to reps, enrolled in sequences and counted as leads.
The result: Submissions that match junk rules you approve are marked and held in a review view: not routed, not enrolled, not counted as leads, and released in one step. Nothing is deleted.
Check whether this job fits
Answer these without sending files or logins. Nothing is submitted unless you choose to contact us.
What you get
- The junk rule sheet
- The hold-marker and view set-up steps
- The seeded test results
- The release steps and the review checklist
Included
- One form path into one CRM portal
- A junk rule sheet you approve, with free-text judgement only for ambiguous cases
- A hold marker and a review view, built by us on a test portal and repeated by your admin from our steps
- A seeded test of 30 junk and 30 genuine synthetic submissions written by your team, 10 of each kept back from us until the run
- A one-step release path and a note on the first 14 days of full review
Not included
- Deleting any submission
- Blocking submissions at the form or the server: that is a separate job
- Real-time blocking of a live attack
- Anything that decides alone about ambiguous cases
- A promise that every bot is caught
How we know it’s done
Agreed with you before work starts. Each check produces evidence you keep.
Your team writes 30 junk and 30 genuine synthetic submissions and keeps 10 of each from us until the run; we write the rules on the other 40. On the 10 junk kept back at least 9 are held, and on all 30 junk at least 28 are held. This is a smoke test on invented submissions: it says nothing certain about live traffic.
Evidence: The 60 submissions, which 20 were kept back, and where each landed.
Of the 30 genuine submissions, none is deleted, and any that is held shows in the review view and is released in one step. A genuine submission that is deleted or held out of the view fails the whole test; 0 lost in 30 shows a loss rate under about 9.5%.
Evidence: The genuine submissions and the review view.
Held items never reach an owner, a sequence or a lead report.
Evidence: The workflow log and a lead report for the test period.
Release restores normal routing for a held test lead.
Evidence: The release test.
Genuine submissions with accents, plus-addresses and single names are not held.
Evidence: The edge-case rows in the test set.
Sign-off. You re-run the test yourself, check the list of severe errors and sign off in writing. Payment follows sign-off.
If it fails. If the agreed pass mark is not met, we say so, hand over the results and the failures, and do not call the output working. You do not pay for this scope; we agree whether to stop or re-quote.
When it fits, and when we stop
It fits when
- You can add a hold property and a saved view to the form's records
- Your plan can stop held records being routed and enrolled
- A named person reviews the held list
We stop and tell you if
- Your plan cannot suppress enrolment or routing for a held record
- Nobody will review the held list
- You want junk deleted automatically
What could go wrong
Release restores normal routing for any held record, and your admin can switch the hold off in one step.
Scroll the table sideways to read it all.
| Risk | How we handle it |
|---|---|
| A real buyer is held and never read | We hold and never delete, the reviewer reads the whole list for the first 14 days, then a sample, and a missed lead is released in one step. |
| A model decides alone on an ambiguous case | A model is used only for ambiguous free text and never decides alone; those cases go to the review view. |
| Bots change their pattern | The rules are a baseline at a date. The monthly service re-runs the seeded set and samples routed leads for junk that got past the hold. |
A separate automated reviewer checks the rules, the seeded results and the release path before hand-over, and no person on our side reviews unless your agreement names one. Your named reviewer reads the held list and releases or keeps each item. 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 junk rules, the hold marker, the review view and the pass marks
- Build the hold on a test portal with notifications off
- Submit your 30 junk and 30 genuine test leads; record where each lands
- Check held items never reach an owner, a sequence or a lead report
- Independent automated review of the rules, the results and the release path
- Hand over the steps; your admin switches it on
This is a one-off job, not a subscription or emergency cover. We confirm eligibility, price, a start window and a delivery date before you accept, and work starts only after agreed inputs and access are in place. Nothing is sent, merged, posted or changed in your live systems by us, and an enquiry creates no charge or booking.
Need to keep it working?
Discuss a monthly check that re-runs the seeded set and samples routed leads for junk that got past the hold.
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
- Hidden fields and challenge tools at the form can stop some bots before they reach the CRM. They may also turn away real people, so test them. news.ycombinator.com
Questions
Do you delete junk?
No. Held items stay in a review view until a person releases or removes them.
Will this stop all bots?
No. It holds submissions that match rules you approve, and the held list is read by a person.
How is a model used?
Only for ambiguous free text, and it never decides alone. Clear junk and clear genuine cases follow your rules.
Send an enquiry
Send us
- The form, the CRM and what a junk submission looks like, with invented examples
- Roughly how many submissions arrive a week and what share is junk
- Who would review the held list
Later, once you agree
- Your 30 junk and 30 genuine test submissions, 10 of each kept back from us until the run
- A test portal or a sandbox with notifications off
- The approved junk rule sheet
- A secure handoff route agreed before any real data is shared
You keep the CRM and the form. We test with invented submissions into your portal with notifications off, and your admin builds the hold. Which service reads your records is agreed in writing before any real record is shared; the entry test uses invented or redacted data, or your own provider account.
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 “crm-fake-form-leads-held-for-review” as the subject.