Guide
Where a recruitment desk actually loses its week
Ask a recruiter where their time goes and they will say sourcing. Watch the desk for a week and it is rarely sourcing. It is the packaging, chasing and re-keying around it.
In short
On a recruitment desk, the activities that feel most like recruitment are the least automatable, and the ones that feel like admin are the most. Duplicate submission checks and submittal packaging are usually worth automating first: both are mechanical, both are expensive to get wrong, and neither requires the software to make a judgement about a person.
Who this is for: Staffing firm owners and delivery managers deciding where automation would actually pay, and what to leave alone.
10 min read · Last reviewed
Automation projects in staffing tend to start at the most visible activity rather than the most expensive one. Sourcing feels like the job, so sourcing gets the tooling budget, and the hours keep disappearing anyway.
This is an attempt at an honest breakdown of where a desk's week goes, and — more usefully — which parts respond to automation and which do not.
We are not going to put a percentage on any of these. We have not measured your desk, and a number invented here would be worth less than the hour you would spend measuring it yourself. The shape of the problem is what transfers between firms; the magnitudes do not.
The six places the week goes
| Activity | What it really is | Automatable? |
|---|---|---|
| Sourcing | Searching, plus re-searching a database nobody has re-mined in a year | Partly — the re-mining much more than the judgement |
| Screening | Reading CVs against a requisition, repeatedly, under time pressure | The reading, yes. The decision, no. |
| Submittal packaging | Reformatting a CV, writing a summary, filling a portal form | Almost entirely |
| Duplicate checks | Confirming this candidate has not already gone to this client | Entirely, and it is the highest-value check on the list |
| Compliance chasing | Collecting documents, tracking expiry, re-collecting | Largely — the chasing, not the verification |
| Timesheets and approvals | Reminding people to submit, reminding others to approve | Almost entirely |
The pattern worth noticing: the activities that feel most like recruitment are the least automatable, and the ones that feel like admin are the most. Which means the honest version of an automation project in staffing gives recruiters back the hours they are least proud of spending.
Duplicate submittals deserve their own paragraph
Of everything on that list, duplicate submission detection is the one with the worst ratio of effort to consequence. The check is trivial. The cost of missing it is a client relationship, and occasionally a contract.
It is also the check most likely to be done from memory, because it depends on knowing what a colleague did three weeks ago. That is precisely the kind of knowledge that does not survive a busy week, a holiday, or a recruiter leaving.
Doing it properly means matching on more than a name. The same candidate arrives as a different file, from a different source, with a different spelling and a different phone format. Matching on normalised email, phone and a fuzzy name comparison, against submittals rather than against the candidate record, catches the cases a name search misses.
Submittal packaging is the quiet one
Nobody lists this when asked what takes their time, and it is frequently the largest single block. A submittal means reformatting the CV to a template, stripping contact details, writing a summary that argues the case, and re-entering fields into a portal that will not accept a paste.
The summary is the part worth being careful about. It is where invented experience enters a process, usually with good intentions — a recruiter rounding up a gap because the candidate is genuinely strong. A generated summary should be assembled from what the CV actually evidences, and should say where evidence is missing rather than quietly filling it in. If there is no measurable outcome in the CV, the right output is a line saying so, not a plausible number.
See it working
Our screening example generates the client-facing summary and advocacy note from cited evidence only. Where the CV states no measurable outcome, the note says the line was left out rather than invented — which is the behaviour that matters when the summary reaches a client.
See a submission summary built from evidenceYour ATS stays your ATS
Almost every firm we talk to has considered replacing its ATS and almost none should. The system holds years of history, every recruiter's habits are built around it, and the migration cost is dominated by data and retraining rather than by licences.
The more productive move is to build around it. Most of what firms actually want — better screening output, automated submittal packaging, duplicate checks, expiry alerts, margin reporting by desk — sits alongside the ATS and reads from it, rather than replacing it. That also keeps the blast radius small: if a piece of automation is wrong, you switch it off and the desk carries on.
What not to automate
- The decision to reject a candidate. Produce the evidence and the reason, route it to a person, and keep the record.
- Client-facing claims about a candidate's experience. Generated text that overstates is worse than no generated text.
- Anything that writes to your candidate database without review. A system that can silently create or merge records can silently destroy history.
- The first conversation with a candidate. This is the part of the job that is actually the job.
Common questions
- What should a staffing firm automate first?
- Usually duplicate submittal detection and submittal packaging, in that order. Duplicate checks have the worst ratio of effort to consequence — trivial to run, expensive to miss — and packaging is often the largest hidden block of time. Both are low-risk because neither makes a judgement about a person.
- Do we need to replace our ATS to automate our workflow?
- Rarely. Most of what firms want sits alongside the ATS and reads from it. Migration costs are dominated by data history and retraining rather than licences, and building around the existing system keeps the blast radius small — if something is wrong you switch it off and the desk carries on.
- Will automation reduce recruiter headcount?
- That is a business decision rather than a technical outcome, and we would be wary of anyone promising it as a benefit. What automation reliably changes is the mix: less time on packaging, chasing and re-keying, more on the conversations that actually place people. Whether that turns into fewer recruiters or more placements per recruiter is yours to decide.
- Is it safe to let a software vendor near our candidate database?
- It depends entirely on whether that vendor might one day compete with you. Ask directly whether they source, submit or place candidates. We do not and will not — we build software for staffing firms and run no desk of our own, which is a deliberate limit on what we will take on.
Related guides
Screening CVs against a brief without a black-box score
Most screening tools answer the wrong question. They tell you how well a CV matches a brief, when what a recruiter needs to know is what the CV actually evidences and what it simply does not say.
Your database is a better source than your job board
Most agencies pay a job board every year to find strangers, while sitting on thousands of people who already know them. The reason that list goes unused is specific, and it is not that matching is hard.
Where to go from here
What we build
For staffing
Software for the firms that place people, built around the ATS you already run — and a company that will never compete with you for a placement.
Rather see it run
Database reactivation
Six decayed ATS records against one live role — what can be recovered from free-text notes, what has rotted, and why one of them is worth a call a job board could never tell you about.
Or skip both — we are a small company in Vadodara that builds exactly this kind of thing, with no obligation and no sales sequence. Get in touch, or try the automation finder, which will tell you when the answer is not to automate. +91 93131 52136 · hello@twishivplatforms.com

