Atlin, a Melbourne-based management consultancy that delivers large ICT and business transformation programs, was running an expression-of-interest campaign across five roles, and had the problem every busy HR team knows: too many applicants per job, arriving faster than anyone could open them. Each role drew roughly 1,300 applicants, 1,347 on the role we confirmed, and about 6,000 across the campaign. Every application was a profile, a CV and a set of answers, readable one at a time and impossible to sort, filter or share the way the team needed.
Working as a delivery partner, with Atlin's approval throughout, we built the HR data pipeline that turned each role's applicants into one clean spreadsheet and one CV folder per role, added the columns the team asked for, and matched every applicant against the team's own screening survey. The whole engagement ran from build start to final delivery in about two weeks.
If you are an HR or operations lead with too many applicants per job and no practical way to compare them, this is what fixing that looks like in practice, including the parts that did not go perfectly.
| Metric | Result |
|---|---|
| Roles in the campaign | 5 |
| Applicants across the campaign | About 6,000 |
| Applicants on a single role | 1,347 |
| Survey responses matched against applicants | 360 |
| Time from build start to final delivery | About 2 weeks |
| First role delivered | Day two of the build |
Too Many Applicants per Job? Why HR Teams Get Stuck
If you have ever posted one role and woken up to hundreds of applications, you already know the problem. Too many applicants per job is not really a volume problem. It is a format problem: applications arrive in a platform designed for reviewing candidates one at a time, while the decisions you need to make, who fits, who is local, who responded to your follow-up, all need them side by side.
That is why one of the most common questions HR teams ask is how to manage job applications on LinkedIn once a role takes off. The answer is rarely a new recruitment platform. It is getting the applicants into a structure you control, then adding the information that turns a long list into a shortlist.
This applies whether you hire for your own business or for clients. Recruitment agencies feel it most, because they also have to send shortlisted candidates on in their own branded CV format, which was exactly the third requirement in this project.
The Brief: Five Roles, One HR Team, No Way to Compare
Atlin's HR team came to us with a list of requirements to streamline their recruitment process. The five roles covered People & Culture, Digital & Artificial Intelligence, Engineering & Integrated Logistics Support, Project & Commercial Management, and Change & Transformation. Three requirements mattered most:
- Get every applicant's details into a format the team could sort and share, per role.
- Keep every CV alongside the data, organised by role.
- Convert candidates' own CVs into Atlin's standard CV template, so shortlisted profiles could go to clients in one consistent format.
There was also a request to tidy up a SharePoint library. We told the team we thought that one could be handled more simply without AI, which is the kind of answer we would rather give at the start than discover halfway through a build.
Can You Export LinkedIn Applicants to Excel? Start With LinkedIn's Own Routes
If you are asking this, start with the routes LinkedIn itself provides, because they keep your account safe:
- Download applicant CVs from the job post itself, which is fine for a handful of candidates.
- Connect LinkedIn to your applicant tracking system. LinkedIn's hiring integrations, Apply Connect and Recruiter System Connect, are designed to move applicants and profiles into your ATS. LinkedIn's own integration guides say these come at no extra cost with a full Recruiter seat, though some ATS vendors charge for their side.
Avoid browser extensions and tools that promise to scrape or "export" applicants for you. LinkedIn's help center lists prohibited software and extensions, including bots and other unauthorised automated methods used to access the service or download data, and LinkedIn can restrict accounts that use them. Your recruiter account is worth more than a shortcut.
Getting the applicants out is only the first step, and usually the smaller one. What decides whether a hiring team saves time is what happens next: structuring every role the same way, adding the columns that matter, matching applicants against your other data, and producing CVs in your own format. That is the part we built for Atlin.
What We Built for Atlin's HR Team
We built a Python data pipeline that took each role's applicants and produced, for every role in the campaign:
- A candidate spreadsheet with every applicant's details in consistent columns.
- A folder of CVs for the same applicants, ready to open or forward.
- 01Collect each role's applicants
Applicant details and CVs for every role in the campaign, with the HR team's approval.
- 02One spreadsheet and CV folder per role
Consistent columns, with the job post and country of residence added on request.
- 03Match against the screening survey
A Completed Survey Y/N column, so the team could see who had engaged.
- 04HR screens and shortlists
The people decisions stayed with the HR team.
- 05Convert shortlisted CVs to Atlin's template
Only for candidates worth progressing, not all of them.
Each role's applicant details and CVs were organised into a spreadsheet and a CV folder. The data was enriched with job post and country of residence, then matched against the team's screening survey to add a Completed Survey column. HR screened and shortlisted, and only shortlisted CVs were converted into Atlin's template.
The first role's spreadsheet and CVs went back to the team on day two of the build. Their response the next morning: "Thank you for your work, this looks great!", followed straight away by two requests that made the data more useful.
The columns HR actually needed
Once the team saw the first spreadsheet, they asked for two additions:
- Country of residence, because the roles were expressions of interest for opportunities in the Middle East and North Africa, and location was the first filter they wanted to apply.
- The job post each applicant came from, so a spreadsheet could be forwarded on its own without losing its context.
- We deliver a real spreadsheetNot a mock-up
- HR uses itOn actual candidates
- HR asks for a columnCountry, job post, survey status
- Added to every rolePast and future deliveries
Both went into every role from then on. This is normal and healthy. The fastest way to find out which columns a team needs is to hand them a real spreadsheet on day two, not to write a specification for a fortnight.
Matching applicants against the screening survey
Atlin had also sent applicants a screening survey through Microsoft Forms, and 360 people had responded. The team asked us to match the survey responses against the applicants and add a simple Completed Survey: Y or N column, so they could see at a glance which candidates had engaged beyond the initial application.
We matched the two data sets and returned the list of applicants who had both applied and completed the survey. That one column turned two disconnected lists into a shortlisting signal.
CV conversion, deliberately held back
We converted sample CVs into Atlin's standard template to prove the approach. Then the team made the right call on scope: convert only the candidates worth progressing, after screening, rather than every applicant. Reformatting more than a thousand CVs for people who would never be shortlisted is cost without value. Organise everything, screen with people, transform only the shortlist.
What Did Not Go Perfectly, and What It Taught Us
A case study that only lists wins is an advert. These are the parts worth knowing before you commission similar work.
At first the role looked empty, because we were looking at its talent pool, not its applicants. Agree the source on day one.
Two separate lists became one shortlisting signal: a Completed Survey Y or N column on every role.
Too many ways of writing PMP, CSM or MBA to clean reliably by machine. We said so rather than ship it.
Cleaning names is harder than it looks. Many applicants had added certifications to their name: PMP, CSM, PSM, ACP, MBA and more. The team asked for clean names only. We tried several approaches, but with so many variations in how people write their credentials, no method removed them reliably every time, and we told the team that rather than ship something that looked clean and was quietly wrong. If a field drives a decision, a human check beats an automated guess.
Be precise about which data you mean. One role appeared to have no applicants. It turned out we were looking at the role's talent pool rather than its applicants, which held 1,347 people. A five-minute conversation fixed it. On any data project, agree at the start exactly which list or report is the source of truth.
Agree data handling before kick-off. Candidate data is personal data. Before any build, settle who provides the data and through which approved route, where files are stored while the work runs, who can see them, and when they are deleted. Settling it on day one also sets the pace of the whole job, because every later step depends on it.
Is This Kind of HR Automation Right for You?
This is a good fit if:
- You run high-volume hiring, where a single role attracts hundreds or thousands of applicants.
- Your team copies candidate details into spreadsheets by hand.
- You need to join applicant data with something else: a survey, an assessment, an internal database or a client's requirements.
- You send shortlisted candidates to clients and need them in one consistent CV format.
If a role attracts hundreds or thousands of applicants and your current system cannot give you clean, structured data, an applicant pipeline is worth scoping, especially if you need to match applicants with a survey or send CVs in one format. If you hire a handful of people a year, or your applicant tracking system already gives you clean data, use what you already have instead.
It is not the right first step if you hire a handful of people a year, or if your applicant tracking system already gives you clean, structured data. In that case the honest answer is a better use of what you already have, not a build.
If you are comparing an HR automation consultant with an agency or an in-house hire, our checklist on how to tell if an AI agency can actually build it gives you the questions to ask, and our plain-English guide to what automation costs a business owner sets out typical budgets.
Whatever you decide, treat applicant data as the personal data it is. Keep it inside your own systems, limit who can access it, and delete what you no longer need once a campaign closes. Our guide to secure document sharing for HR and recruiting covers how to move CVs and candidate files without emailing them around.
What This Would Cost a Person to Do by Hand
The fair comparison is not "automation versus nothing". It is automation versus the recruitment coordinator who would otherwise do it. Here is that comparison, with every assumption on the table so you can swap in your own.
The assumptions. Handling one applicant by hand means opening the application, downloading and filing the CV, copying the details into a spreadsheet and checking them against the survey list. We estimate 3 minutes per applicant, and show 2 and 4 minutes as the fast and slow cases. For cost, we use an Australian recruitment coordinator on AU$65,000 a year, between Payscale's average of about AU$55,000 and the AU$73,000 to AU$83,000 midpoints of advertised roles, working the standard 38-hour week, plus the 12% super guarantee. That comes to about AU$37 an hour, before recruitment fees, equipment or management time.
For this campaign alone. About 6,000 applicants at 3 minutes each is roughly 300 hours: close to eight working weeks of one full-time coordinator, or about AU$11,000 in wages, with a range of AU$7,400 to AU$14,800 depending on speed. And that is eight weeks during which the first candidates have been waiting, which in a competitive market is its own cost.
Why reuse changes the maths. The manual cost repeats in full for every new role. The pipeline does not: once it exists, structuring, enriching and matching a new role's applicants is a re-run, with a person spending around 15 minutes per role starting it and spot-checking the output.
| Volume | Applicants | By hand (3 min each) | Wage cost by hand | Pipeline, human checks only |
|---|---|---|---|---|
| This campaign | ~6,000 | ~300 hours (about 8 weeks) | ~AU$11,100 | ~1.25 hours |
| 50 roles at ~1,300 each | ~65,000 | ~3,250 hours (about 1.6 full-time years) | ~AU$120,000 | ~12.5 hours (~AU$460) |
| 60 roles at ~1,300 each | ~78,000 | ~3,900 hours (about 2 full-time years) | ~AU$144,000 | ~15 hours (~AU$560) |
| 50 roles at a lighter 250 each | ~12,500 | ~625 hours (about 16 weeks) | ~AU$23,000 | ~12.5 hours |
| 60 roles at a lighter 250 each | ~15,000 | ~750 hours (about 20 weeks) | ~AU$28,000 | ~15 hours |
The pipeline itself is a one-off build, starting from our US$5,000 fixed-scope pilot. At Atlin's volume, the wage cost of doing a single campaign by hand is already larger than that, and across 50 to 60 roles the manual route costs over AU$100,000 a year in coordinator time against a few hours of checking. Even at a lighter 250 applicants per role, it is more than 600 hours of work a year that a person no longer has to do.
Two honest caveats. These are estimates, so run your own figures through our hire versus automate calculator. And a pipeline needs occasional maintenance when a source system changes its format, which belongs in the budget, though it is hours, not months. The bigger saving is not the wages. It is that your coordinator spends those weeks talking to candidates instead of copying them into a spreadsheet.
What HR and Recruitment Process Automation Costs
A focused applicant pipeline like this one fits within our fixed-scope pilot, starting at $5,000, because it is one workflow with a clear definition of done: every applicant, every CV, the right columns. Adding CV conversion to your own template, automated screening against your criteria, or a pipeline that runs for every new role moves it into a larger build, priced before you commit.
For a figure for your own hiring process, our automation quote tool gives a ballpark in about a minute, the hire versus automate calculator compares a build with adding a recruitment coordinator, and our pricing page covers ongoing support. Our AI automation development service explains how a pilot is scoped.
How This Fits With the Rest of Your Hiring Stack
A clean applicant pipeline is usually the first step, not the last. Once applicants are in a consistent structure, the same data can feed:
- Automated resume screening. Once the data is clean, you can automate resume screening against your own criteria, with people making the final call. We cover where AI document processing is reliable, and where it is not, in our guide to AI document processing.
- Workflow automation that routes shortlisted candidates to hiring managers and updates your tracker, which we explain in AI agents for business automation.
- The right tool for each step. Integrations, AI and plain code each suit different jobs, and our comparison of RPA versus AI agents sets out where each belongs.
The same organise-first approach is behind our other data case studies: getting a full patient registry out of a clinical system with no export button, and the complete AI agent system we built for a 40-doctor network. Before any of it, you can run the 90-second systems access test on your own tools.
The Competitor Pulse Check
| Factor | How we approached it | A typical approach |
|---|---|---|
| Starting point | Organise every applicant into one structure first | Start with AI scoring on whatever data is easy to reach |
| First delivery | A real spreadsheet on day two, refined with the team | A specification phase before anything is delivered |
| Use of AI | Only where it adds value; we said when something did not need AI | AI applied to every requirement |
| CV conversion | Shortlist only, after human screening | Every applicant, whether or not they progress |
| Data quality | Told the team when a cleaning step was not fully reliable | Ships output that looks clean |
Frequently Asked Questions
Can you export LinkedIn applicants to Excel?
Start with LinkedIn's own routes: download CVs from the job post for small numbers, or connect LinkedIn to your applicant tracking system through LinkedIn's hiring integrations, which LinkedIn says come at no extra cost with a full Recruiter seat. Avoid third-party scraping or "exporter" extensions, which LinkedIn's rules prohibit. Once the data is in your hands, a pipeline can organise it into one spreadsheet per role.
What can I do when a job post gets too many applicants?
Get the applicants into one structured list per role, then add the columns that matter for your decision, such as location or whether they completed your screening survey. In this project each role drew roughly 1,300 applicants, about 6,000 in total, and a sortable spreadsheet plus a folder of CVs per role let the team filter and compare candidates instead of reviewing profiles one by one.
Can AI automatically screen resumes?
AI can sort and pre-screen resumes against criteria you define, but it works best once applicants are in a clean structure, and the final decision should stay with people. In this project Atlin's HR team screened and shortlisted themselves, and only shortlisted CVs were converted into their template.
Is AI automation useful for recruitment agencies?
Yes, mostly for the repetitive data work around candidates rather than the hiring decision itself: organising applicants, matching them against other lists, and converting CVs into your agency's own branded template for clients. Keeping people in charge of screening is what makes it safe and accurate.
How much does it cost to process job applications manually?
At about 3 minutes per applicant, roughly 6,000 applicants takes around 300 hours, close to eight working weeks of a full-time recruitment coordinator, or about AU$11,000 in wages at AU$65,000 a year plus super. Across 50 to 60 high-volume roles the manual cost rises to well over AU$100,000 a year, while a reusable pipeline needs only around 15 minutes of human checking per role.
How long does it take to automate an HR data pipeline like this?
This engagement took about two weeks from build start to final delivery, with the first role delivered on day two. Most of the time went into refining columns with the team and matching the survey data, not into building the pipeline itself.
How should HR teams handle applicant data during automation?
Treat it as personal data. Agree before the build who provides it and through which approved route, keep it inside your own systems, limit who can access it, and delete what you no longer need once the campaign closes.
How much does recruitment process automation cost?
A focused applicant pipeline like this fits within our fixed-scope pilot starting at $5,000. Adding CV conversion, automated screening or a pipeline that runs for every new role makes it a larger build, which we price before you commit.
Want This for Your Hiring Process?
If your team is copying candidates into spreadsheets by hand, or sitting on hundreds of applications it cannot sort, we can tell you in one call whether a pipeline like this is worth building for you, and what it would cost. Book a free strategy call and bring your current process, not a specification.