From 5 days to 90 minutes: the case for automating candidate sourcing

Most recruitment agencies have the same bottleneck. It isn't winning new roles, it isn't closing candidates, it isn't even the rate the market pays. It's the days between a fresh job spec hitting a recruiter's desk and a usable shortlist being ready to show the client.

In the agencies we work with, that gap is typically three to five working days. By the time the shortlist is together, the strongest candidates have already taken offers from a competitor who got there first.

Walk a recruiter through their week and the maths becomes hard to argue with. A junior consultant on £35k costs the agency around £165 a day fully loaded. Spending 60% of that on manual sourcing — running the same boolean searches in LinkedIn Sales Navigator, downloading CVs from Reed, cross-referencing against the internal database — costs roughly £100 a day, per recruiter, before anything is closed. Across a 12-person desk, that's £1,200 a day, £24,000 a month, going on a task that doesn't directly produce revenue.

This is the case for automating the sourcing layer.

When we say a SunEdge AI build delivers a shortlist in 90 minutes, that doesn't mean the system replaces recruiters or runs without supervision. It means this:

A recruiter pastes a job description into the dashboard. The system extracts must-haves, salary, location, seniority. Within 90 minutes, the same dashboard shows a ranked shortlist of 20–25 candidates, sourced from LinkedIn, Reed, Indeed, CV-Library, Totaljobs, Monster, inbound applications, and the agency's internal candidate pool. Every candidate has an AI-generated score, a breakdown of why they fit, and a personalised outreach draft ready to send.

The recruiter still does the work that requires judgement — picking who to contact first, deciding which candidates to push to the client, having the actual conversations. What they stop doing is the work a script can do better: running the same query in seven tools, opening profiles one at a time, copying data into spreadsheets.

Sourcing automation works best for roles where the requirements are knowable from a CV — backend engineers, finance directors, account executives. It works less well for senior leadership where networks and references matter more than profile data, and for roles where culture fit is the primary screen.

For those, the automation still helps — it just plays a smaller role. The system surfaces a longer list of candidates worth a conversation, and the recruiter does the conversation. The time-saving is smaller, maybe 50% rather than 80%, but it still compounds across a desk.

The real argument for automating sourcing isn't speed, it's reallocation. Every hour your recruiters claim back from manual sourcing is an hour spent talking to candidates, briefing clients, closing offers. Those are the activities that produce revenue, and they're the activities recruiters are actually good at. Automation isn't about replacing the work — it's about taking the part you'd never miss off their plate.

If you want to see what 90 minutes looks like end-to-end, the demo on the home page shows the full pipeline running. It's not a hypothetical — it's the exact workflow we build for every client.