AI Automation
Thread Recruitment
An automation that turned two to three hours of daily data entry into something that runs itself, and that Chris can adjust without us
A recruitment founder was copying candidate answers from his social media campaigns into his HR system by hand. We scoped it in two calls, laid out his options, and built something he can change himself without a developer.
~NZ$30,000 of founder time back per year
Two to three hours a day of copying records is roughly 600 hours a year. Valued conservatively at NZ$50 an hour, that is about NZ$30,000 of founder time returned to actually placing candidates. It adds up further every time Chris changes a campaign, because there is no developer to brief and no invoice to wait on.
The problem
What was actually broken
Thread Recruitment fills high-volume roles by running social media ad campaigns. Candidates answer a series of questions in a chat, and every answer needed to end up in TeamTailor, the HR system the business runs on. The only thing connecting the two was Christopher.
- Two to three hours a day copying candidate answers from ManyChat into TeamTailor by hand.
- High campaign volume meant the backlog grew faster than he could clear it.
- Every manual entry was a chance for a typo to skew the reports the business relies on.
- Time spent on data entry instead of placing candidates.
Why it mattered now
Christopher had already spoken with 18 engineers trying to get this solved. Every day without a fix was another few hours lost and another chance for his two systems to disagree.
The approach
How we approached it
Christopher had been pitched at by a lot of engineers already, so the priority was a short, clear process that ended with him choosing the approach rather than being sold one.
- 1
Two one-hour calls
We spent two hours in total understanding the tools he used, the candidate information he collected, and exactly where the manual work happened. The problem did not need a long engagement, so we did not run one.
- 2
Lay out the options, let him choose
We came back with a shortlist of ways to solve it, each with its trade-offs in plain English. Christopher picked the one that fit his business and his budget.
- 3
Build it so he can change it
The brief we agreed on was not just to stop the copying. It was to make sure Christopher could change what gets captured later without needing a developer.
- 4
Test it on real candidates
We built it and ran real candidate applications through it until it was reliable enough to leave running on its own.
The solution
What we built
Candidate answers now land in TeamTailor the moment someone finishes the chat. The part Christopher values most is that he runs it himself.
New questions, set up by Chris
Christopher adds a question to his ManyChat flow and names it following a simple pattern. The name tells the system where the answer belongs in TeamTailor, so he can set up a new question himself in minutes.
Every answer in the right place
Each answer goes to the right spot: the candidate's profile, a note, or a specific questionnaire. One naming pattern covers all three.
Candidates added instantly
When a candidate finishes the chat, they are created or updated in TeamTailor straight away, with no manual step.
Checked before it is saved
Every application is checked and cleaned up before it reaches TeamTailor, so bad data never gets into the HR system.
Before & after
From manual to automatic
Before
After
Candidate answers read out of a ManyChat conversation and retyped into TeamTailor, one field at a time.
The moment a candidate finishes the chat, they appear in TeamTailor with every answer filled in.
Changing what gets captured meant briefing a developer and waiting for them to make the change.
Chris names a new question in his ManyChat flow and the answer lands in the right TeamTailor field. No developer needed.
Manual retyping meant typos, inconsistent records, and reports nobody fully trusted.
Every application checked before it is saved. Over 500 processed cleanly.
I spoke with 18 engineers before finding Anthony. He was the first who made me feel heard and could translate complex technical ideas into plain English. He delivered exactly what he promised: a seamless integration that just works. The manual work is gone, our data is clean, and communication throughout was top-tier.
The impact
What it changed
The manual work is gone, the data is clean, and Christopher can adjust it himself as his campaigns evolve.
Return on investment
Two to three hours back, every working day.
That is the better part of a working day returned each week, with cleaner data as a bonus. Because Christopher makes changes himself, it keeps paying off every time a campaign changes rather than needing another build.
More work
Related case studies
The Data Lab
My Marketing Agency built every client's performance report by hand each month, pulling GA4, Meta and Shopify numbers into spreadsheets and slide decks. We built them The Data Lab, a live reporting platform with AI recommendations they now share with every client.
Southern Rivers Fly Fishing
A sole operator was losing 10 to 20 hours a week to admin spread across Google Calendar, Gmail and Xero invoices done by hand. We built a system that does the behind the scenes work and leaves him email replies ready to send.
About us
We turn your goals into AI and software that actually works
A team of product engineers based in Queenstown, NZ. We work with you to understand the problem first, then build the right thing — not just the possible thing.
Book a consultation