A weird thing is happening inside sales and marketing teams.
They are still hiring people, obviously.
But they are also starting to hire for a different kind of leverage.
Not just “Who can do the work?” More like: “Who can design the system that gets the work done by humans and agents together?”
That shift is early, but it is very real. Microsoft’s 2025 Work Trend Index says 45% of leaders see expanding capacity with digital labor as a top priority over the next 12 to 18 months, 78% are considering hiring for AI-specific roles, and nearly half of leaders say their companies are already using agents to fully automate workflows or processes in some areas. Salesforce’s 2026 State of Sales announcement adds that 54% of sellers have already used agents, while Highspot’s 2025 enablement research says AI is becoming the foundation of modern GTM motion.
That does not mean humans are gone.
It means role design is changing.
The old hiring reflex
When teams get overloaded, they usually ask:
do we need another SDR?
do we need another campaign manager?
do we need another coordinator?
do we need another ops person?
Sometimes the answer is yes.
But more and more, the better question is:
which repetitive work should remain human?
which repetitive work should be assisted?
which repetitive work should be agent-operated with human review?
That is a much sharper way to think.
The role audit I would run
Before opening a req, I would audit the current work.
Use a table like this:
Task | Frequency | Judgment required | AI can assist? | AI can operate? | Human owner |
|---|---|---|---|---|---|
Prospect research brief | daily | medium | yes | yes, with review | SDR / AE |
First-draft follow-up email | daily | low-medium | yes | yes, with approval | AE |
Final deal strategy for late-stage enterprise opp | weekly | high | yes | no | manager / AE |
Webinar recap and repurposing | weekly | medium | yes | yes | marketing |
CRM field cleanup and routing | daily | low | yes | yes | RevOps |
Pricing-page messaging decisions | monthly | high | yes | no | product marketing |
This forces you to separate three kinds of work:
human-only
human-led with AI assist
human-supervised, agent-operated
That is an incredibly useful distinction.
The new GTM roles I expect to grow
Not every company will use these exact titles, but the functions are coming fast:
agent builder / workflow designer
someone who turns repetitive GTM work into reliable automationsAI enablement lead
someone who trains teams on where AI helps and where it clearly should notknowledge / prompt librarian
someone who organizes approved messaging, proof, and process instructions so outputs stay usefulROI analyst for AI initiatives
someone who measures whether the automation actually saved time or improved outcomes
These are not science-fiction jobs.
They are very practical jobs created by practical mess.
A simple pilot for a small team
You do not need a reorg to start acting like an AI-native GTM team.
I would run one 30-day pilot.
Week 1: pick two workflows
Choose two that are repetitive and painful.
Examples:
inbound lead triage
prospect research briefs
webinar repurposing
event follow-up
call-summary routing
renewal risk flagging
Week 2: define the guardrails
Write down:
what the workflow should do
what inputs it needs
what good output looks like
when a human must review
what should never be automated
That last one matters a lot.
Week 3: assign an owner
Every AI workflow needs a human owner.
Not “the ops team.” Not “marketing.” A person.
If nobody owns the workflow, nobody improves it.
Week 4: measure the result
I would track:
time saved
error rate
cycle-time reduction
adoption by team
downstream business result
If the workflow saves time but creates garbage, it is not leverage.
It is a more efficient mess.
A concrete example
Let’s say your SDR team spends hours building account briefs.
You do not necessarily need another SDR first.
You may need one person who can design a workflow that:
pulls firmographic context
summarizes recent company changes
surfaces likely pain points
drafts a signal-based angle
routes the brief into the rep’s workspace
That single workflow might give your current team more lift than another headcount line.
That is why I think “agent builder” becomes a real GTM advantage.
My founder take
The best teams will not divide into “AI people” and “normal people.”
They will become mixed teams where humans set direction, use judgment, and build trust while agents handle more of the repetitive operating layer.
That changes hiring.
It also changes management.
Because soon, a strong GTM operator will not just manage people.
They will manage systems, prompts, workflows, and digital teammates too.
What I’d do next week
Before opening the next GTM req, I would run a one-hour role audit.
Not to avoid hiring.
To make sure I am hiring the right kind of leverage.
Sometimes the next best teammate is a person.
And sometimes it is the person who knows how to build the teammate who never sleeps.
