A CMO can appoint an AI lead inside marketing and still find every practical change waiting on that one person. The partner team needs a campaign adapted. Demand generation needs account research. Communications needs a monitor it can rely on. One employee becomes the department’s translator, builder and fixer.

Build on the change already underway.

Two published surveys, Jasper’s and Norwest’s, find the same thing. Many teams already have a named AI owner, and self-trained marketers are a major source of the AI already in use. Ask each functional head to name the one person on their team who has already automated something without being asked. That list is your first cohort, and it is usually longer than the CMO expects.

Q16

How far has the marketing job actually changed?

Roles and measures of performance changed for people-and-agent work12%
69 responses
New builder or AI-operations roles were hired or assigned15%
89 responses
Existing employees received practical workflow training23%
132 responses
Teams have access to courses or self-directed learning28%
163 responses
Roles and development plans have not changed18%
105 responses
Don’t know3%
19 responses
577 people answered. All listed answers, including don’t know, stay in the total. Full readout and cross-tabs ↗
Jasper, State of AI in Marketing 2026Norwest, B2B benchmark 2025

Replace one generic training session with a real assignment. A partner marketer brings a recurring co-marketing request. They set out the proof it needs, try a method, spot the weak result and fix the instructions. A systems person from marketing ops or IT helps with connections and access. The partner lead judges whether the asset is usable. The marketer learns the work they will actually own.

Keep the specialist’s judgment in the process.

Routine creative is a good example. A designer approves the templates, the allowed variations and what gets rejected. A generalist uses that to produce event banners or social variants. The design team gets two things back. Fewer banner requests, and the right to pull the template if quality slips. Say that trade out loud, or the designers will read the change as headcount pressure.

What is actually in place · Q15

How far has routine creative moved beyond the design queue?

Generalists produce routine creative work from designer-approved templates and quality rules.

Not happening 20%A few one-off trials 30%Used in several recurring jobs 28%The normal way of working, reviewed regularly 17%Don’t know 5%
581 people were asked; each row counts only the people who answered that row. Each row adds to 100% before rounding. “Used in several recurring jobs” and “the normal way of working” are shown as separate answers, not added together. Answer wording is shortened here; the questionnaire and explorer carry the full text. Explore every response ↗

A suggested way to produce routine event assets

01

Creative director

Approves templates, layout rules and examples of unacceptable work.

02

Partner or demand marketer

Creates approved variants with the right copy, partner logos and sizes.

03

Design review

Checks the exceptions, spot-checks the routine work and tightens the standard where the same error keeps coming back.

New concepts and sensitive communications still need a designer’s direct judgment.

The organizational question is who sets the quality bar and who is allowed to produce which assets. Judge the arrangement by turnaround, how often a designer had to step in and how many defects got through, not by assuming more generated assets justify a smaller design team.

Canva, design tools that AI assistants can use

Give the agent a job description too.

Give the monitor the same three things you would give a contractor. A job description, which finding 04.2 spells out. A budget line, so its running cost is visible. A kill date, on which it is retired unless someone can show they acted on its output in the last month. An agent nobody reads is a subscription with extra steps. Name the person who can change or stop it.

Block’s Buzz, a shared workspace for people and AI agents

Ask each functional head to describe these arrangements on their own. If the AI lead sees the normal way of working while the team sees a few one-off trials, look at a recent task together. That conversation will point to the next assignment, training need or permission change. Give the action to a named person and review the actual work after thirty days. When the board asks what AI has done for capacity, that reviewed task is the answer to bring, with the before and after figures.

Bring this discussion to your GTM team.

The team benchmark and workshop connects the operating questions with your own workflows, priorities and people.

Request the team benchmark and workshop ↗