How to Build AI Into a Marketing Team’s Workflow, Approvals Included

For marketing and agency leaders who want AI to save time across the whole process, not just in one step.

I’ve led a change of this kind before. At BBDO Gulf, I led the integration of digital campaigns into the agency’s traditional marketing framework, with planning and budgeting that worked across disciplines, in a department of 60 people across several GCC markets. That experience is why I look at AI as a change to how work moves through a team, and only then as a set of tools.

What does workflow redesign mean in marketing?

Workflow redesign means changing the steps, roles and approvals of a piece of recurring work so that a new capability, in this case AI, actually saves time across the whole process. It answers three questions for each step: who does it, what AI prepares and who signs it off.

McKinsey’s 2026 State of AI survey found that nearly three-quarters of AI high performers have fundamentally redesigned their workflows around AI, up from 55% a year earlier. Among other respondents, the share is about a quarter (McKinsey, August 2026). Marketing leaders seem to know this. In Gartner’s 2026 CMO Spend Survey, 70% said their internal marketing processes aren’t mature enough to implement and scale AI (Gartner, May 2026).

Adding tools vs redesigning the workflow

	Adding AI tools to the old process	Redesigning the workflow
Drafting	Faster	Faster
Roles	Unchanged and unclear	Written down step by step
Approvals	Same sequence, more drafts waiting	Consolidated and scaled to risk
Quality	Some people over-trust output, others redo it	A person owns each final decision
New tasks	Nobody owns them	Assigned to named people
Time saved overall	Often lost at the next bottleneck	Visible across the whole process

Start with one recurring workflow

Redesigning everything at once rarely works. Choose one workflow your team runs every week or month, such as the social content calendar, campaign assets or monthly reporting. If you’ve already mapped how your team works, pick the one with the most waiting and rework.

Decide who does what at each step

For each step, agree on what AI prepares and what a person decides. A common split looks like this:

Step	AI prepares	A person decides
Brief	A structured first draft from meeting notes	Objective, audience and message
Copy	Draft options and variations	Which option, the edits and the voice
Design and video	Variations and resizing	Brand standards and the final call
Review	A pre-check against the brand checklist	Approval of claims and risk
Reporting	The data, pulled together	What it means and what to do next

Write this down. When the split stays implicit, people either over-trust the output or quietly redo all of it, and the time savings disappear.

Fix approvals before they become the bottleneck

When drafting gets faster, review becomes the slowest step. A few changes help:

• Define what each reviewer checks. Brand reviews brand, legal reviews claims, the client lead reviews the message.
• Combine feedback into one consolidated round instead of several sequential ones.
• Agree on turnaround times for each reviewer.
• Use a short brand checklist before work goes to review, so reviewers aren’t catching basic errors.
• Match the level of review to the level of risk. A social caption and a regulated product claim shouldn’t follow the same route.

Give the new work an owner

Some tasks appear that didn’t exist before: keeping a library of prompts and templates up to date, checking AI output against brand guidelines, tracking what has changed. Assign them to someone. Unowned work is the first to slip.

Training matters here too. People need practice with the new split of work, which I wrote about in what AI enablement means for a marketing team.

How to redesign a workflow, step by step

1. Choose one recurring workflow with visible waiting or rework.
2. Record how long it takes today, stage by stage.
3. Write the “AI prepares / a person decides” split for each step.
4. Redesign the approval route: who checks what, in one round, within an agreed time.
5. Assign owners for the new tasks.
6. Train the people involved on the new way of working.
7. Run it for a month, compare against the starting numbers, then move to the next workflow.

Common mistakes to avoid

• Redesigning on paper only. If the people who do the work weren’t in the room, the new process will drift back to the old one.
• Speeding up drafting and nothing else. Faster drafts with the same approval route means a longer queue.
• Sending every piece through the same review. High-risk content needs more checks; low-risk content needs fewer.
• Assuming AI output is ready to publish. Every piece still needs a person who owns the final version.
• Changing several workflows at the same time. It becomes hard to tell what helped and what didn’t.

Agencies vs in-house teams

The approach is the same, but the pressure points differ. In an agency, the client’s approval round is often the longest wait, so agreeing on review steps with the client is part of the redesign. In an in-house team, the waits are more often internal: legal, product or leadership sign-off. In both cases, the work is to make it clear who checks what, and by when.

Where I can help

I work with marketing teams and agencies on workflow redesign: mapping the work, agreeing who does what and fixing the approval route so the time AI saves actually reaches the calendar. The rules that go with it are covered in a generative AI policy for marketing teams.

Final thought

Choose one workflow, map it with the people who do it and redesign the approval step first. It’s usually where the most time is hiding.

Frequently asked questions

Sources

1. McKinsey & Company, “The state of AI in 2026: On the road to ROI,” August 25, 2026. Read on mckinsey.com
2. Gartner, 2026 CMO Spend Survey, press release, May 11, 2026. Read on gartner.com

Related articles