Roles we recruit
Seven practices, 27 functions. Start at the practice when you cannot yet name the title; go straight to the function when you can.
Browse all 229 roles Brief us on a seatFlat 15% of first-year base · shortlist in 10 working days
Twelve seats sit under AI and marketing automation, and most of the titles are newer than the tools they use. We split them by the step from process map to production each one owns, screen for workflows still running months after launch rather than demos, and charge a flat 15% only once your hire starts.
Titles in this function are new and inconsistent: one company’s AI engineer is another’s Zapier specialist. We start with what you need built and where it has to run, then find the seat.
From process to production, and the seat that owns each step
Moves: the right processes chosen to automate
1 seat
Mapping how work is done today, choosing which steps to automate, and designing each workflow with the human checkpoints it needs.
Brief us when teams want AI everywhere and nobody has decided which processes are worth automating first.
Moves: reliability, security and cost at scale
2 seats
The architecture for AI in the business: models, data access, retrieval, security, cost controls and how the systems fit together.
Brief us when several teams build AI workflows separately, and you need one design before it becomes a maintenance problem.
Standards for automation across the business: platforms, naming, error handling, monitoring and who can build what.
Brief us when automations have multiplied, nobody knows which ones are running, and customers find the failures first.
Moves: no-code workflows shipped and maintained
4 seats
Day-to-day automation across marketing and sales tools: triggers, workflows, integrations and the fixes when something breaks.
Brief us when your teams have a backlog of manual tasks and nobody whose job it is to automate them.
Building and hosting workflows on n8n, including custom nodes, self-hosting, code steps and AI agent flows.
Brief us when you run or plan to run n8n, often self-hosted for data control, and need someone fluent in it.
Building and maintaining Zapier automations across your tools, including multi-step zaps, filters and AI steps.
Brief us when your teams live in Zapier and need someone to build properly, document and clean up what already exists.
Building scenarios on Make, including routers, error handling, data stores and AI modules, for complex multi-step workflows.
Brief us when your automations run on Make and have outgrown what the team can maintain alongside their day jobs.
Moves: custom agents and integrations in production
2 seats
Custom automations that use language models and APIs: lead enrichment, research agents, data cleanup and integrations, written in code where no-code stops.
Brief us when your workflows need logic, scale or integrations that no-code platforms cannot handle reliably.
Reliable, monitored workflows across systems: APIs, webhooks, queues, retries and logging, built to keep running.
Brief us when your automations fail silently, or every integration is a one-off that only its builder understands.
Moves: output quality and review time
2 seats
The instructions, examples and evaluations that make model output consistent: prompt libraries, test sets and the measures of quality.
Brief us when your AI outputs vary from excellent to unusable, and nobody is testing them systematically.
Hire a Prompt EngineerContent production with AI in the loop: workflows, brand and accuracy review, and an output volume the content team can stand behind.
Brief us when you want more content without lowering quality, or AI drafts are reaching customers without review.
Moves: customer adoption of AI features
1 seat
Helping customers adopt your AI features: technical discovery, solution design, demos and proofs of concept with AI at the center.
Brief us when your product sells AI capabilities and buyers need a technical expert to show it working on their data.
Measured on
AI automation attracts people who build impressive prototypes. We screen for what happens after launch: workflows still in production, error handling and maintenance, data and security judgment, and whether the business actually adopted what they built.
We ask for one automation they shipped that is still in production: what it replaced, how often it fails, how they know, and who maintains it. Demos with no production history are where we dig.
Candidates build or fix a small workflow on your stack, whether n8n, Zapier, Make or code, inside our screen. We watch how they handle errors and edge cases, not just the happy path.
We ask how they handled personal data, CRM write access, API keys, rate limits and human review. An automation with full write access and no review is a risk, not a feature.
We check with the team that used it whether the workflow changed how they work. An automation nobody trusts gets switched off within a quarter.
Not sure which seat you need? Tell us the process you want to automate, the tools involved and who will maintain it, and we will tell you which of the twelve owns it.
Public US benchmarks, read October 2026. Most of these titles are too new for salary surveys, so these are the closest published benchmarks. Engineers who build custom AI systems usually sit above them.
Each bar runs from the 25th to the 75th percentile, the middle half of reported pay.
Base salaryMedian base salaryAverage total pay, base plus variable
Automation Engineer
Built In, self-reported, all automation engineers
$107,000 median base$119,141 average total
Marketing Automation Manager
Salary.com base benchmark
$113,573 to $137,157 base
Salesforce Administrator
Salary.com base benchmark, a neighboring seat
$81,358 to $96,711 base
Automation Engineer: Built In median base salary and average total compensation, self-reported, all automation engineers, reviewed October 1, 2026. Marketing Automation Manager and Salesforce Administrator: Salary.com base salary, 25th to 75th percentile, United States, as of October 1, 2026. No public benchmark yet covers n8n, Zapier or Make specialists or AI workflow designers as separate titles. Public aggregates, not ZaphyrRecruit placement data.
The clock starts when the brief is signed and stops when three to five calibrated candidates are in front of you. Most of that time is spent talking to people, not searching for them.
Day 1
Forty-five minutes with whoever the hire reports to. We leave with the number the seat owns, the range you can pay and the two or three things that genuinely disqualify.
Days 2 to 7
We build the list from people doing the job now, approach them directly and screen every name against the brief before it reaches you.
Days 8 to 10
Three to five people, each with a page: what they have run, why they fit this seat, pay expectations, notice period and where they are weak.
After the shortlist
Interviews run at your pace. We handle scheduling, references and the counter-offer conversation, and nothing is payable until your hire starts.
Three clients, quoted as they wrote to us about searches we ran for them. We publish no names or logos without permission.
They build the automations that take manual work out of sales, marketing and operations: enriching and routing leads, researching accounts, cleaning CRM data, drafting first versions of content and reports, and connecting tools that do not talk to each other. Good ones design for failure too, with logging, retries and human review wherever the output reaches customers.
Start from the work, not the title. If your workflows connect common SaaS tools with simple logic, a Zapier, Make or n8n specialist ships faster and costs less. If you need custom logic, high volume, your own data in a model, or integrations no platform covers, you need an engineer who writes code. Many teams start with a specialist and add an engineer later.
A flat 15% of the hire’s first-year base salary, payable only once they start. On a $110,000 base that is $16,500. There is no retainer, no exclusivity and nothing to pay if the search does not produce a hire.
Public data is thin because the titles are new. Built In’s self-reported figures put automation engineers at a $107,000 median base with average total compensation of $119,141 (October 2026), and Salary.com puts marketing automation managers at $113,573 to $137,157 base. Engineers who build custom AI systems usually sit above both.
It can be, with the right design: scoped permissions, logging, validation before writes, rate limits and human review for anything customer-facing. We test candidates on exactly this judgment, because a workflow with full write access and no checks can corrupt data faster than a team can fix it.
Ten working days from a signed brief to a shortlist of three to five calibrated candidates. Interviews and the offer then run at your pace. A short build or debugging exercise on your platform sits inside our screen, so your interviews can focus on your processes.
Yes. An automation engineer, a CRM administrator and a RevOps manager hired as one structure is how we run a pod, briefed together so later hires do not restart the search. The same 15% applies to each hire, payable as each one starts.
Tell us the process, the tools involved and who will maintain what gets built. We reply within one business day with which of the twelve seats it is and what it pays.
Most of our automation searches are filled from people we already know, the week an employer briefs us. Set up a job alert for operations and AI roles, or send your CV and we will come to you when a seat opens.
AI & Marketing Automation sits in our AI-Native GTM practice, next to the newer roles it often grows into and close to the operations teams whose systems it automates.