Firms are more and more hiring for this new function.
xAI is hiring a Head of GTM, Techniques and Brokers.
Notion is hiring a GTM AI and Innovation Supervisor.
Zapier introduced on a Director of GTM Innovation.
Whereas the titles themselves range from firm to firm, If you happen to learn the job descriptions what they’re all hiring is an Agent Operator.
Each GTM workforce is about to want one. Most groups have already got a shadow model, normally a high-agency RevOps lead, a curious AE, or the marketer who turned their private Claude workflow into the workforce’s. They’re doing the job with out the title, the price range, or the instruments to do it effectively. That’s altering, with extra corporations hiring for Agent Operators.
Why this function, why now
The Agent Operator is a job more and more being employed for as a result of six issues occurred on the identical time, and any considered one of them by itself would’ve been a footnote. Collectively, they reshape the org chart.
Six forces converged on the identical time, and the function needed to emerge
1. Brokers went from chatbots to coworkers
A yr in the past, “AI agent” seemed very totally different than it does now. The work brokers can do has expanded roughly 10x in 18 months. The work people have to do to deploy them has expanded with it. Defining the duty, evaluating the output, dealing with the sting circumstances, optimizing the prompts: none of that goes away while you swap a human for an agent, it simply strikes to whoever’s working the agent.
2. The tooling stack hit manufacturing grade
At this time, there’s an actual stack behind agentic GTM. Clay for knowledge and orchestration, n8n and Gumloop for workflow plumbing, Lindy for vertical brokers, inside orchestration on high of Claude, GPT, and Gemini. The instruments graduated from “attention-grabbing” to “manufacturing.” Manufacturing stacks want homeowners, the identical approach each different manufacturing system within the firm has one. No person runs a Salesforce occasion with out an admin. No person runs a advertising and marketing automation platform with out an ops lead. And but most GTM groups are working a fleet of brokers with no person whose precise job is to ensure they maintain working.
3. The perimeters are saturated
The following layer of leverage is the transfer from a thousand personal experiments to a handful of deployed, measured, repeatable programs that the entire workforce runs on. That requires any individual whose precise job is to make that translation occur, taking the IC’s intelligent immediate and turning it into infrastructure your entire workforce advantages from. With out that individual, you don’t have an AI technique.
4. The work itself modified form
Work that was once fulfilled by workers is now agent work. However agent work isn’t free. Somebody has to outline the duty in writing, consider whether or not the output is nice, deal with the circumstances the agent can’t, and optimize the system over time. With out that individual, you could have a fleet of brokers and nobody watching the street. The work didn’t disappear. It moved upstream, and it concentrated. Which suggests the leverage of the individual doing it concentrated too.
5. The headcount math flipped
“Do extra with much less” has been a typical line for years. Most groups have heard that as a hiring freeze. It isn’t, precisely although as a result of it’s extra like a hiring freeze for the outdated org chart.
The brand new math is people plus brokers, and the mathematics solely works if somebody is aware of how you can deploy and supervise the brokers. An Agent Operator isn’t a price line, it’s the multiplier on each different GTM rent you make. With out one, your three-person SDR workforce has three folks. With one, your three-person SDR workforce has three folks and a fleet of brokers doing the analysis, checklist constructing, and first-touch outreach behind them. The Operator’s worth is measured by how a lot they will enhance everybody else’s output.
6. The client expects 1:1 at scale
Generic outbound is useless. Consumers can spot a templated sequence from 100 yards, and reply charges have collapsed for anybody nonetheless attempting to scale the outdated approach. The bar is now customized, researched, contextual outreach, at quantity.
You possibly can’t hit that bar with people alone. The unit economics don’t work. You possibly can’t hit it with brokers alone both. They miss nuance, hallucinate context, and produce output that’s assured however flawed if no person’s watching. The one path by means of is people plus brokers, stitched collectively cleanly, with an Agent Operator working the seam. The seam is the function.
Agent work, then and now
If the function nonetheless feels summary, the best option to floor it’s to have a look at how the underlying work modified. Two years in the past, an “AI agent” was a characteristic. At this time it’s a coworker. That shift is what makes the Agent Operator function load-bearing.
What ‘AI agent’ meant 18 months in the past vs what it means right this moment, and why that shift calls for a brand new sort of supervision
The 18-months-ago column isn’t a strawman. We’ve all sat by means of demos of single-turn chatbots that deflected an FAQ, received celebrated as innovation, and received deserted six months later when no person wished to take care of them. They have been options, not programs, and so they didn’t want an proprietor as a result of they didn’t actually do something.
The 2026 column is structurally totally different. Multi-step autonomous workflows that stay contained in the system of report, run repeatedly, and produce output that downstream people depend upon, are programs. Techniques have failure modes. Techniques drift. Techniques degrade silently if no person’s watching them. The explanation most agent deployments quietly fail in month three isn’t that the know-how stopped working. It’s that the know-how saved working whereas the world round it modified, and there was no person whose job was to note.
What the Agent Operator truly does
Strip the title politics out of the dialog and the work breaks down into 4 phases, run on a loop. Each Agent Operator job description ought to map to those 4.
4 phases, run on a loop. Every one feeds the subsequent. With out the loop, you could have brokers and nobody watching the street.
Outline
The Agent Operator writes the job description for each agent on the workforce. Not the immediate, the JD. What’s the duty, in writing? What’s a profitable output appear to be? What inputs does the agent get? What does it escalate? What does it by no means do, ever, even when requested?
Free job definition produces unfastened output, the identical approach a unfastened hiring temporary produces a unfastened rent. The Agent Operator’s first transfer on any new agent is forcing the precision that prompts alone by no means demand.
Deploy
The Agent Operator picks the stack. The choice is “which device is finest for this particular job, given our knowledge, our workforce, and the half-life of the know-how.” It’s tremendous nuanced to every firm.
The pic isn’t everlasting.. Stacks have a 6-9 month half-life proper now. The Agent Operator’s job is making the subsequent swap painless when the higher device reveals up, which it should.
Consider
The Agent Operator builds the evals: a pattern set of inputs, an outlined rubric for what “appropriate” seems to be like, a option to rating outputs at scale, and a baseline efficiency quantity that will get tracked over time.
When the underlying mannequin will get up to date, evals catch the regression earlier than prospects do. When a brand new immediate will get pushed, evals inform you whether or not it’s truly higher or simply feels higher. With out evals, you’re managing on vibes. With evals, you’re managing the way in which a gross sales supervisor manages a rep ramping up: towards a measurable bar that doesn’t transfer primarily based on whose flip it was to demo the device that week.
Optimize
Brokers don’t keep the identical, fashions replace, instruments deprecate options, consumers change habits, inside priorities shift. The Agent Operator iterates on prompts, context, knowledge sources, and guardrails on a daily cadence (weekly for high-volume brokers, month-to-month for lower-volume ones), not when one thing breaks.
That is the half that compounds. A workforce working iteration cycles each week for six months has a fleet of brokers that’s measurably higher than a workforce working the identical brokers on autopilot. The brokers on autopilot degrade as a result of the world round them modified and no person adjusted.
The three issues an Agent Operator does not personal: the technique itself (that’s the GTM lead’s job), the shopper relationships (nonetheless people), and the price range (RevOps or finance). They’re a pressure multiplier on these capabilities, not a substitute for them.
The market’s already pricing this in
If you happen to suppose this thesis sounds early, the labor market disagrees. The function exists, it’s already being employed, and the one factor the market hasn’t executed but is decide on what to name it. The titles are inconsistent, however the work is similar. We imagine the perfect identify for this new function is the Agent Operator.
The labor market is transferring sooner than the org charts. The function exists. The titles are nonetheless being found out.
Based on Apollo, GTM engineering postings grew roughly 205% throughout 2025, from round 1,400 in mid-2025 to over 3,000 by January 2026. That’s two consecutive years of doubling, and the curve continues to be steepening. Throughout the 200 fastest-growing corporations tracked by GTMnow, “GTM Ops, Techniques, or Engineering” titles outnumber “Advertising and marketing Engineer” titles by greater than 15 to 1. The operate is consolidated, and the titles haven’t caught up.
The historic analog right here is RevOps in 2018. In 2017, “RevOps” was a distinct segment title utilized by a handful of corporations, largely in SaaS. By 2019, it was a board-level dialog. By 2021, each Sequence B firm had a head of RevOps. The operate consolidated as a result of the work grew to become obligatory, the headcount math demanded it, and some high-profile hires gave the remainder of the market permission to observe. The Agent Operator function is on roughly the identical curve, besides the cycle is going on sooner as a result of the underlying know-how strikes sooner.
Methods to construct this function inside your workforce in 30 days
The sensible place to begin isn’t a elaborate job posting. It’s the minimal viable model of the function that will get began this quarter with somebody already in your workforce.
Days 1 to five: Audit what’s already deployed.
Pull an entire checklist of each AI device, agent, automation, and GPT subscription anybody on the GTM workforce is utilizing. The audit will floor two issues: a shadow stack a lot greater than the management workforce thought existed, and one or two people who find themselves clearly already doing elements of the Agent Operator job with out the title.
Days 6 to 10: Decide the individual.
The very best Agent Operators come from considered one of three locations: (i) a powerful RevOps background plus real technical curiosity; (ii) an engineering or PM background that pivoted into GTM; or (iii) a high-performing IC who turned their private AI workflow into the workforce’s. The individual you need is normally the one already working unofficial brokers on their very own time. Speak to them. Carve out 50% of their present function for the primary 90 days. Don’t make it a full function but, earn the headcount with outcomes.
Days 11 to twenty: Decide three high-leverage brokers to standardize.
Don’t attempt to take over every thing without delay. Decide the three workflows that produce essentially the most repeated work and have the clearest measurable output. Lead analysis, post-call summaries, and outbound personalization are frequent beginning factors. Standardize every one: write the duty definition, decide the device, construct the eval, set the exception path. Get these three working clear earlier than including a fourth.
Days 21 to 25: Arrange the measurement.
Construct the KPI dashboard earlier than the function goes stay. Agent-attributed pipeline, time-to-deploy for brand new brokers, eval rating development, headcount equal saved. The Agent Operator’s personal efficiency evaluation at day 90 ought to run off this dashboard. So ought to be the case for changing their function to a full-time rent.
Days 26 to 30: Run the primary evaluation cycle.
Pull a pattern of each agent’s output. Learn it. Rating it towards the eval. Discover one factor to repair. Repair it. Doc what you modified and why. Ship the subsequent iteration. Then do all of it once more subsequent week.
By day 30, you could have an individual, a stack, three deployed brokers, a measurement system, and one full iteration loop accomplished. That’s the proof of idea. From there, the function compounds. By day 90, you could have the case for changing it right into a everlasting rent and the info to again the headcount dialog.
When AI execution compresses, supervision turns into vital
There’s a model of this story that sounds purely tactical. Job descriptions, eval frameworks, KPI dashboards. That framing misses the larger transfer.
What GTM groups are literally constructing after they rent an Agent Operator is a supervision functionality. When software program creation compresses, distribution turns into the constraint (the Cursor lesson). When AI execution compresses, supervision turns into the constraint. The groups that personal the supervision layer personal the output benefit. The groups that don’t are working on stale brokers, drifted prompts, and a variety of confidence about an AI technique that’s quietly producing mediocre work.
The explanation this function compounds is that each different GTM rent improves when an Agent Operator is doing the job effectively. The SDR’s analysis is best. The AE’s name prep is quicker. The marketer’s campaigns ship sooner. The CSM’s renewals are hotter. The Agent Operator isn’t changing any of these folks. They’re making every of these folks materially simpler, each week, on a curve that compounds as a result of the brokers themselves get higher with iteration.
The Agent Operator at a look.
Tag @GTMnow so we are able to see your takeaways and assist amplify them.
ElevenLabs crossed $500M ARR, up from $350M at finish of 2025 in simply 4 months. The Sequence D brings in BlackRock, Wellington, Nvidia, Santander, and 30+ creatives together with Jamie Foxx and Eva Longoria. Voice is changing into the first interface layer for enterprise AI, and the most important cash is beginning to agree.
How high GTM leaders keep within the hiring course of. Vikas Bhambri joined Andy Mowat’s Whispered Hiring podcast to interrupt down the method he runs personally. For instance, he provides candidates suggestions within the final quarter-hour and watch how they take it.
GTM 189: AI Will Make 90% of Govt Selections Inside 2 Years | Steve Lucas, Boomi CEO
Pay attention by means of the hyperlinks within the web page above or by looking wherever you get your podcasts “The GTMnow Podcast.”
Village – raised $9.5M led by Upfront Ventures, with participation from GTMfund, to repair specialty pediatric care. Practically 1 in 5 U.S. youngsters face developmental, behavioral, or psychological well being wants and most households are left quarterbacking fragmented suppliers alone. Village brings the care workforce collectively in a single place.
Hightouch – raised $150M at a $2.75B valuation to construct the agentic advertising and marketing platform: AI brokers that run in your precise buyer knowledge to analysis audiences, generate inventive, and execute campaigns 24/7.
-
Account Success Supervisor, SMB at Crossbeam (Distant – Boston / New York / Philadelphia)
-
Founding Account Govt at Atrix (Hybrid – New York, NY)
-
Principal Gross sales Engineer at Vividly (Distant – US)
-
Director of Buyer Success at Fastbreak AI (Charlotte, NC)
-
VP, Buyer Success (EMEA) at Author (Hybrid – London, UK)
See extra high GTM jobs on the GTMfund Job Board.
Upcoming occasions you received’t need to miss:
-
SaaStr Annual: Could 12–14, 2026 (San Mateo, CA)
-
GTMfund Dinner: Could 14, 2026 (San Francisco, CA)
-
RevStar Summit: June 3, 2026 (Toronto, Canada)
-
GTMfund Dinner: June 9, 2026 (London, UK)
-
Dreamforce 2026: September 15–17, 2026 (San Francisco, CA)
-
INBOUND: September 16–18, 2026 (Boston, MA)
-
Pavilion GTM2026: September 28–October 1, 2026 (NYC, NY)
-
CVC Week by Counterpart Ventures: September 29, 2026 (San Francisco, CA)
-
Buyer Success Week: October 5-9, 2026 (NYC, NY)
-
TechCrunch DISRUPT: October 13–15, 2026 (San Francisco, CA)
Some GTMnow Community love to shut it out – we respect you.
