There are two very different things people mean by recruitment automation. One is on the candidate side: parsing CVs, scheduling interviews, sending status updates. The other is on the client side: finding companies that are hiring, getting in touch, following up and booking calls.
Most agencies have at least tried the first. The second is where the time is quietly disappearing. This post is about that second one, and about being honest on where automation helps and where it makes things worse.
The client-side workload nobody counts
Try this. Next time your team does business development, write down every step, however small. Something like this comes out:
- Scan job boards and news for companies that might be hiring.
- Decide which ones are worth contacting.
- Find the right person at each company.
- Look up and double-check their email.
- Write a first message.
- Send it.
- Remember to follow up in four days. Then again.
- Watch for replies.
- Reply, qualify, book a call.
- Log everything somewhere.
That’s ten steps for one company. Do it for fifty companies a week and it’s a full-time job. Most of those steps are repetitive, rule-based, and don’t need a human’s judgment. That’s what makes them good candidates for automation.
What’s safe to automate
Signal tracking. Watching job postings, funding news, leadership changes and expansion announcements against your niche. A person can’t realistically watch this daily. Software can.
Company filtering. Applying your rules (size, region, sector, “already works with agencies”) to throw out poor fits before anyone spends time on them.
Contact finding and verification. Identifying the likely hiring owner and checking that the email is valid. Done well, this protects your sender reputation as well as saving hours.
First-draft messages. Using signal details (the role, the funding round, the new leader) to prepare a relevant first draft. This is a draft, not a finished product. More on that below.
Follow-up scheduling. Sending the second and third message on time, and stopping the moment someone replies. People forget this part. Systems don’t.
Reply routing and reporting. Telling you who replied and what they said, and turning the week’s activity into five simple numbers.
What to keep human
Anything that needs judgment about a relationship. If a past client replies, or someone refers you in, a person should handle it.
Approving the voice. Automated messages should still sound like you. Someone with taste should approve the wording before it goes out, and review it again when results change.
The conversation itself. Once a decision-maker replies and wants to talk, that’s a human job. The call is where trust is built, where you ask good questions, and where you decide whether the role is actually one you can fill.
Pricing and terms. Don’t automate negotiations. Ever.
The mistakes that give automation a bad name
When agencies say “we tried automation and it didn’t work”, it’s usually one of these:
- Blasting a big list with one generic message. That’s not automation, that’s spam with extra steps. Relevance comes from signals and tailoring.
- Skipping verification. Bounces damage your sending domain, and then even good emails land in spam.
- No follow-up discipline. One email isn’t a sequence, and endless follow-ups are harassment. Pick a sensible cadence and stick to it.
- Never looking at the numbers. If nobody reviews replies and calls booked each week, nobody notices when something quietly breaks.
- Set and forget. Messages go stale. Targets change. Someone has to look after the system.
A sensible way to start
You don’t need to automate everything on day one. A good order is:
- Start with signals and filtering. It’s the biggest time saver and it’s low risk, because nothing is sent to anyone yet.
- Add contact verification. Your sender reputation is worth protecting from the start.
- Move to tailored first drafts, with approval. You see every message type before it goes live.
- Add follow-ups. Once you trust the first message, a short follow-up sequence is the natural next step (see our follow-up sequence workflow).
- Add reporting. The weekly five numbers described in measuring outbound results.
Each stage should run for a few weeks before you add the next. That way you know what each change did.
What this looks like in practice
A team using automation well spends its week differently. Instead of scanning job boards and verifying emails, the founder looks at a short report on Monday: who was contacted, who replied, which calls are booked. They spend their time on the calls and the candidates, which is where an agency’s skill actually lives.
If you want an example of a workflow that’s built this way from end to end, see recruitment outreach to new leaders, which is one of the signal-led approaches that suits automation well.
Do you build it or buy it?
You can wire this together yourself with a handful of tools, if you have someone technical and a few weeks to spare. The catch is the ongoing care: workflows break, data providers change, messages need rewriting. Many agencies find it cheaper to have the system built and run for them than to hire and manage an automation engineer plus a business development team to operate it. Either way, the principle is the same: let software do the repeatable steps and keep your people for the conversations.
The short version
Automate the finding, filtering, verifying, drafting, follow-up timing and reporting. Keep people on approval, relationships, calls and pricing. Start small, watch the numbers, and add one piece at a time.