One recruiter. Two operators.
A specialist recruitment desk in equipment hire, OEM, and industrial: a tight vertical where the same 268 companies keep showing up in the work. We embedded Otto to run the operations that sit around the recruiting.
The problem was not ability. It was bandwidth.
One recruiter runs the desk, and runs it well: knowing the market, knowing the candidates, building client trust over years. The vertical is narrow and relationship-driven.
Around that sits the operational load every small desk knows: checking job boards for signals across dozens of companies, by hand, on random days. Writing cold approaches from scratch. Scanning LinkedIn most mornings hoping to catch a hire announcement or a growth signal. Trying to find an hour for a LinkedIn post before the week gets away.
None of it is hard work. All of it eats time that should go to recruiting. The desk was doing good work, when it could get to it.
Not another tool. A parallel operator.
The decision was not to subscribe to another SaaS product or automate one task. It was to build an embedded AI operations layer: a single agent, Otto, running the recurring, research-heavy, output-generating work the business needs but a human does not need to do.
The distinction matters. A tool helps you write better emails. Otto wakes at 1am, scans the entire target market, and has a ranked briefing waiting before coffee. It is not a feature. It is a parallel operator.
Phase 1 was scoped to four automations, built and deployed over several weeks, on the desk’s own workflows and target company list.
Four automations. One audit.
Runs every night at 1am. Scans all 268 target companies for new listings, active hiring, team growth, and emerging gaps. By 6am a briefing is waiting: which companies posted overnight, what the roles are, and a suggested action: call now, make an approach, or keep watching.
One click on a candidate profile. Otto pulls role type, skills, location, and seniority, searches live roles, scores all 268 companies for fit, and drafts personalised outreach for the top five: real role titles where they exist, honest neutral openers where they do not. Under two minutes, ready to review and send.
Runs Sunday night. Researches what is moving in recruitment, workforce planning, and the industrial sector, then drafts three LinkedIn posts (Monday, Wednesday, Friday), each with a hook, a developed body, and a niche translation for equipment hire and OEM. A weekly feedback form feeds the next plan, so the content improves.
A Monday-morning summary: who was floated, to which companies, and current status. Context for the week. One less thing to reconstruct.
Otto audited all 949 contacts in the database: 124 out-of-scope companies identified, 28 name and email mismatches flagged, 8 missing emails caught. Realistically a full day of someone’s time.
What changed.
Before
- Job boards checked by hand, on random days.
- Cold approaches written from scratch, one at a time.
- Mornings spent scanning LinkedIn for signals.
- Floating a candidate felt like a half-day commitment.
- Content was inconsistent or stressful.
After
- A ranked briefing waits in the inbox by 6am, every day.
- Outreach is drafted from live market data, ready for review.
- The reactive scan is replaced with an actionable summary.
- A float runs in the length of a coffee break, so it happens.
- Three niche-relevant drafts arrive every Monday.
Four entries from the record.
Before the brief existed, the market was checked manually, on the days it was remembered. Now the full 268-company sweep runs every night without fail, and the morning starts with a decision, not a search.
The research behind a float (live roles, company fit, who is hiring right now) used to take two to three hours of desk time. It now takes under two minutes of Otto’s.
Work that stops being expensive stops being deprioritised.
Each week’s posts are graded through a simple form. Otto reads the feedback before writing the next plan. The drafts are not generic recruitment content. Each one answers what a trend means for someone hiring in equipment hire or OEM.
949 contacts checked. 124 out-of-scope companies. 28 mismatches. 8 missing emails. The kind of task that is always worth doing and never gets done, finished as a side effect of the engagement.
The numbers will be the run log’s, not ours.
The structural record is above: companies scanned, runs completed, contacts audited. Outcome figures against the pre-Otto baseline are being measured now.
We publish verified numbers or none. Figures from this engagement are available on request once the measurement period closes.