Skip to content

A SetpointHQ service

The recruitment marketing engineer.

A recruitment marketing engineer builds AI agents and automated systems that attract, source and convert candidates, rather than running each channel by hand. SetpointHQ builds and runs this stack for recruitment agencies and in-house talent teams. Here is what it looks like in practice.

Book a discovery call Or send a message first
The agent stack
  • Hiring trigger agentWatching
  • Candidate language agentListening
  • Answer slot agentScanning
  • Advert testing agentTesting

4 agents, one engineer. Built once, supervised daily, improving weekly. Each is defined below.

The claim

The most valuable recruitment marketer changes with every technology wave. This wave belongs to the one who builds the system.

Recruitment marketing has always minted a new most valuable operator when the technology shifts: the storyteller, then the channel owner, then the funnel optimiser. The agent wave mints the engineer: the recruitment marketer who builds AI systems that attract, source and convert, then supervises them instead of doing the work by hand.

SetpointHQ has been building and running these systems through 2026, on its own pipeline first. The name is new. The service is already running. This page defines the role and sets out how to buy the output.

The lineage

4 waves. 4 most valuable marketers.

Each wave stacked a new layer on the last one’s skills rather than replacing them. The engineer still needs everything underneath, which is why ours started at the bottom of the stack, back in print and broadcast.

  1. 01

    The traditional marketer

    Print and broadcast

    Made people care with story. Reach was bought and attention was earned.

    In recruitment: the employer brand storyteller. Culture, values, the reason to join.

  2. 02

    The digital marketer

    Search and social

    Owned each new channel as it arrived and learned its rules before anyone else.

    In recruitment: the channel operator. Job boards, PPC, programmatic, careers site SEO.

  3. 03

    The growth marketer

    Product-led growth

    Lived in loops, testing and retention. Data replaced instinct as the tiebreaker.

    In recruitment: the funnel optimiser. Source-of-hire data, conversion rates, candidate experience.

  4. 04

    The marketing engineer

    AI agents

    Builds AI agents that run the whole system, and improve it, without being asked.

    In recruitment: the recruitment marketing engineer. Builds the agents that attract, source and convert.

The work

The 4 agents we build.

Each one replaces a repeating manual task with a system that runs continuously. The engineering is choosing what they watch, what good output looks like, and when a human takes over. Every build is scoped to your desks, your sectors and your data.

  1. 01

    The hiring trigger agent

    Watches for signals that a company is about to need help: new job adverts in the roles you fill, funding announcements, a competitor taking a beating in reviews, a fresh incorporation at Companies House. Enriches the contact and drafts the outreach tied to that exact trigger the moment it fires.

    For agencies this is business development on autopilot. For in-house teams, the same machinery points at talent pools instead of clients.

  2. 02

    The candidate language agent

    Pulls candidate feedback, review-site language and drop-off reasons on a schedule, extracts the exact words candidates use to describe the job and its problems, and feeds that language back into adverts, careers pages and outreach.

    Adverts written in the candidate’s words outperform adverts written in the employer’s.

  3. 03

    The answer slot agent

    Finds the questions candidates and clients are asking that nobody in your sector has answered, checks who currently holds the answer in search and in AI assistants, and drafts the page that wins the slot.

    Rankings are earned slowly through authority. Answer slots are won by being the first crawlable page to state the number, the definition or the price.

  4. 04

    The advert testing agent

    Takes one role, generates advert variants across different angles, publishes them, measures response, and retires anything underperforming without being asked.

    Only the winners keep spending your budget and your candidates’ attention.

The loop

Every agent runs the same loop.

Watch, score, draft, approve, learn. The approve step is yours by design: nothing sends, publishes or spends without a human seeing it first.

What the agents learn from your approvals is what makes the stack yours rather than a template. 6 months in, an agent trained on your desk drafts things a generic tool never could.

Watch Score Draft Youapprove Learn Continuous. Supervised, not unattended.

The shift

Same desk. Different physics.

Nothing on the left disappears. It moves to the right, runs continuously, and comes back as a briefing instead of a to-do list.

Before : run by hand

The desk works for the admin

  • Client research1 hour per company
  • BD prospectingMonday morning list building
  • Advert copy2 hours per role, from blank
  • Advert performanceReviewed when things go quiet
  • Candidate outreachBatch and blast

Every hour above is an hour not spent recruiting.

After : run by agents

The admin works for the desk

  • Client researchBriefed and scored in 60 seconds
  • BD prospectingTriggers watched, outreach drafted on fire
  • Advert copy11 asset playbook per job
  • Advert performanceTested continuously, losers retired
  • Candidate outreachDrafted in the candidate’s language

The recruiter recruits. The agents feed the desk.

Not theory

These systems wrote parts of this page.

We run this stack on our own pipeline every day, which is the difference between selling a methodology and selling a working system. A daily engine scans care regulator registrations, Companies House incorporations and live job adverts, scores the signals and drafts the outreach. The SetpointHQ Index measures how 4 AI assistants describe 100 UK employer brands every month, which is how unclaimed answer slots get found. Studio generates the adverts, outreach and playbooks for the desk itself. Your build starts from systems that already work, not a blank page.

See The Index

Hiring trigger agent : this morning

3 new care providers registered in your patch overnight. 2 are already advertising nights. Outreach drafted for both, tied to the adverts they posted.

Answer slot agent : this week

Nobody in your sector holds the answer for “average cost per application care assistant UK”. A page stating the number would take the slot. Draft ready for review.

Advert testing agent : day 6

Variant B (pay stated, shift pattern first) is outperforming variant A on applications. Variant C retired. Budget moved to B.

Who buys this

Two buyers. One stack. Different agents first.

  1. Agency owners and directors

    The 360 desk, with the surround running itself

    Which companies to call this morning, what to say to them, which candidate to pitch at which live role, and adverts that improve without a Friday afternoon rewrite session. The recruiter recruits. The agents feed the desk. Usual first build: the hiring trigger agent, because it pays for the engagement in won clients.

  2. Heads of TA and HR Directors

    Attraction stops being a queue

    Built for high-volume, candidate-short sectors: care, health, logistics, hospitality, construction. Triggers, language, content and testing run continuously in the background, and the team’s time moves to the judgement calls. Usual first build: the candidate language and advert testing agents, because cost per application drops first.

The scarce half

The engineering is the accessible half. Agent tooling gets easier every quarter, and most of it no longer needs code. The scarce half is the recruitment marketing judgement: knowing which trigger matters, which words convert, which slot is worth winning and what good looks like when the agent drafts it.

An engineer without that judgement automates the wrong things faster. Which is why this service is a recruitment marketer who learned to build, not a development shop that learned some recruitment. The years of channel work underneath are what your agents get built from.

100UK employer brands tracked monthly in The Index
4AI assistants measured: ChatGPT, Claude, Perplexity, Gemini
11Assets Studio generates per job, across 6 stages
0Sends, publishes or spends without a human approving it

Worked examples : illustrative

What the engagement looks like inside 2 recruitment businesses.

Scenario 1 : agency

A 6 desk temp agency, health and social care

  • Hiring trigger agent watching care regulator registrations, incorporations and live adverts across their patch
  • Client research brief generated and scored for every trigger that fires
  • Pitch drafts matching available candidates to roles spotted on the radar
  • Advert variants running on the 2 hardest desks
The Monday briefing, once the stack is live

3 providers registered in the patch overnight. 2 are already advertising nights. Outreach drafted for both, tied to the adverts they posted.

An available nurse fits a live role spotted this morning at one of them. Spec angle and pitch message drafted.

Advert test, day 5: the variant with pay and shift pattern first is ahead on applications. Losing variant retired.

What changes: the Monday list-building session stops existing. BD starts every morning already drafted, and the first hour of the day is calls, not research.

Scenario 2 : in-house

A care provider, 40 services, a chronic care assistant shortage

  • Candidate language agent mining review sites, exit feedback and application drop-off reasons
  • Advert testing across the highest volume roles, week in, week out
  • Answer-slot pages drafted for the pay and rota questions candidates ask search engines and AI assistants
  • Trigger agent pointed at talent pools: local closures, competitor instability, seasonal movers
The weekly review, once the stack is live

Language report: “every other weekend” dominates the negative rota mentions this month. Adverts updated to state the rota plainly, drafts ready to approve.

Answer slot: no local page states care assistant pay for the area. Draft page ready. It states the number.

Advert test: stating paid travel time up front is outperforming the generic benefits list. Budget moved.

What changes: attraction stops being a queue of requisitions. The team’s week starts with decisions to approve, not adverts to write.

Both scenarios are illustrative. The mechanics are the ones we run daily; the businesses are composites, because client work is confidential and we do not publish invented results.

How to buy it

It starts with a conversation. Or run the productised version yourself.

Step 1 : the conversation 20 minutes

Tell us about your desks

Your sectors, your data, the pain points. We listen, then diagnose the right first step: sometimes a mapped audit, often a bespoke build scope. Straight answers either way. Step 2 : the build Scoped for you

The engineering engagement

We build the agents your situation calls for, on your data and your desks, train your team to supervise them, and stay on to run and improve them if you want us to. Self-serve route £149/month per seat

Studio

The productised version for hands-on recruiters. One job in, an 11 asset playbook out, plus Desk Radar for live hiring signals. No engagement needed.

Sometimes the right first step is a full map rather than a build. Where that's the case, the Recruitment Marketing Audit covers 7 areas including AI visibility, and we'll say so on the call.

Questions

Common questions

What is a recruitment marketing engineer?

A recruitment marketing engineer builds AI agents and automated systems that attract, source and convert candidates, rather than running each channel by hand. The typical systems watch hiring triggers, mine candidate language, win answer slots in search and AI assistants, and test advert variants continuously.

What does the engineering engagement include?

Building the agents your situation calls for, on your data, your desks and your sectors: typically some combination of hiring trigger watching, candidate language mining, answer-slot content and advert testing. It includes training your team to supervise the systems, and an option for us to stay on and run them. Every engagement is scoped on a discovery call because no two desks need the same stack.

How is it different from a recruitment marketer?

Same judgement, different delivery. A recruitment marketer runs the channels: adverts, campaigns, careers content, sourcing support. A recruitment marketing engineer builds systems that run those channels continuously and improve them from the results. The judgement about what converts is identical, which is why this service is delivered by a recruitment marketer who builds, not a development shop.

Does our team need to be able to code?

No. The systems are built for supervision, not maintenance: your team approves outputs, adjusts what the agents watch and escalates edge cases. Anything that needs engineering time is our side of the engagement. Handover depth is agreed at scoping, from full independence to us running everything.

Which sectors do you work in?

High-volume, candidate-short, transient-workforce sectors: care and health first, then logistics, warehousing, hospitality and construction. The trigger sources, candidate language and compliance constraints in these sectors are where the 20 years of recruitment marketing behind the service were spent.

Where should we start?

2 routes. If you want the systems built for you, start with the £999 Recruitment Marketing Audit: it maps where agents and automation would pay back first, and the engineering engagement builds from that map. If you are a hands-on recruiter who wants to run the productised version yourself, Studio is £149 a month per seat.

Next step

Hire the output, not the headcount.

A recruitment marketing engineer on the payroll is a scarce, expensive hire. The engagement gives you the output without the search: systems built by the practitioner who runs the ones on this page, scoped against your desks in a 20-minute conversation. Hands-on recruiters can run the productised version in Studio from £149 a month per seat.

Book a discovery call Send a message