Marketers Can Finally Win Their Time Back

A while back, I came across a remark from Julia White, the CMO of AWS, that was strikingly blunt. The gist: storytelling and building human connections will always be the most important things in marketing.

Then she added the second half: everything else has changed.

I stared at that line for a long time. Because it's true.

Marketing Has Become a Technical Job

Think about how marketers have lived through the past few years.

First you learned search engine optimization. How to pick keywords, how to build backlinks, how Google ranks pages — a skill set people honed for over a decade until they knew it cold.

Then one day you woke up and found the world had moved on.

Now, for more and more customers, the first touchpoint isn't a search box at all — it's an AI assistant. They ask directly: which of these two products is right for me? Is this vendor any good? The AI researches for them, compares for them, recommends for them.

It's a researcher, a consultant, and a gatekeeper all at once.

What does that mean? It means your content has to be machine-readable, not just human-readable. Whether your positioning is clear, whether your information is structured, whether your signals are trustworthy — these used to be bonus points. Now they're the price of admission.

Julia White says AWS has already observed the shift: customers arriving through AI channels engage more deeply and are higher quality. Fewer casual browsers, more people showing up with a specific question.

For marketers, that's actually good news.

What Is Agent-Based Marketing?

Where to start? Let's start with the first wave.

The first wave of generative AI in marketing handled content chores: writing emails, posting, summarizing, editing copy. Useful, sure — but frankly, it was still "humans aim, AI fires."

The second wave is different. This one is called agents (Agent).

What's an agent? It's a large language model with hands and feet — able to run processes on its own. You give one instruction, and it logs into systems by itself, gathers materials by itself, checks its own work, and delivers the finished job.

Here's an example. AWS needs to localize content into 16 languages. How did that used to work? Machine translation produced a rough draft, then linguists spent two or three weeks fixing grammar and filling in context. Now, agents fix the grammar and add local relevance first, and humans handle only the final gate: making the content genuinely better.

The human job shifted from "fixing the machine's mistakes" to "making human judgment calls."

That's where AI truly creates value. No sci-fi vision required — start with a common bottleneck, and make it faster and better.

An Assembly Line, Collapsed into One Instruction

An even fiercer example: building web pages.

AWS produces roughly ten thousand web pages a year. What was the old process? The marketing team filed a request, the content team wrote, the agency built, then review, publish, QA — each step handing off to the next, moving forward like an assembly line.

Now a marketer says one sentence in plain language: make me a page about such-and-such product. The agent logs into the content management system on its own, assembles the materials, applies the template, adds images and links, handles both search and AI optimization — it even checks the backend rendering for you.

A linear assembly line has collapsed into a single agent workflow.

Remember this lesson: the biggest gains always come from redesigning the process around AI — not sprinkling a little AI over the old process and praying for a miracle.

The marketer's role changed with it. No more assembling assets, chasing approvals, checking links. The questions you answer now are worth more: What story should we actually tell? Why should customers care? How are we different?

Let the machines run the process; keep the judgment for humans. Seen that way, marketing actually becomes more human.

Applause Broke Out in the Conference Room

Of course, mention AI and creative people get anxious: if AI can write, draw, and analyze, what do you need marketers for?

Julia White's estimate stings: a marketing team probably spends only 20% of its time on storytelling, ideation, and connecting with people. The other 80% gets consumed by mechanical work — assembling, verifying, formatting, reporting.

So the right mindset is: hand the friction to AI, keep the judgment for yourself.

She told one detail I especially love. When AWS demoed the new agent-driven web page workflow to its marketing team, spontaneous applause broke out in the room.

Why clap? Because everyone knew: the most annoying, least interesting work — finally, they wouldn't have to do it themselves.

Applause. Actual applause. Have you ever seen a marketing team applaud an internal tool? What does that tell you? The moment people embrace AI is the moment it takes over the work you hate — not the moment it threatens the work you love.

The Era of Dashboards Is Over

One more AWS case, about data.

The old situation was textbook: to see data, marketers had to queue up and wait for the BI team to build dashboards. Inside AWS there were roughly 1,500 dashboards piled up, with another 2,000 requests waiting in line behind them. Everyone wanted data, and the data team had become the bottleneck.

Then AWS built an agent system called AIRO. Marketers no longer request dashboards — they just ask: why is this campaign underperforming? The system pulls together the data warehouse, analysis results, and causal research, explains it to you, and suggests what to do next.

What's the difference? A traditional dashboard tells you what happened. This thing tells you why it happened, and what to do next.

From a reporting tool into a thinking partner.

Of course, humans still need to be in the room. Challenging conclusions, checking logic, making the business call — those are still your job. But the work gets meatier.

The Lesson of 56 Agents

At this point you might be thinking: fine, I get the logic — but where do I start?

Julia White learned this the hard way. She threw down a challenge to her team: become the most AI-forward marketing team in the company. Everyone got fired up and started building tools on their own. After a while, she went and looked at the internal wiki:

56 content agents.

Her reaction: we do not need 56 ways to make content.

That's the warning for every organization. Enthusiasm without coordination quickly turns into duplicated effort and piles of what she calls "AI slop." Letting everyone experiment is fine at the start — but it's not a long-term model for transformation.

How did AWS rein it in? They converged on five core workflows — content is one, with lead management and campaign management among the others. Get the people aligned, and the processes can genuinely be redesigned.

And there's a homespun tactic any company can copy: have employees submit their own "paper cuts" — those maddening little chores they never want to do again. She put it plainly: send us the thing you never want to do again; if you never have to do it again, you'll be happy.

Start with what everyone already knows is slow and painful — not with a grand strategy document. Solve one pain point, earn one measure of trust; solve another, build momentum.

The Agents' Personal Computer Moment

So what's next? Julia White thinks this agent wave is still very early.

She used an analogy: today's agents are like personal computers in the early days — if you wanted a computer, you bought the parts and assembled the machine yourself. Gradually, they'll become plug-and-play, and marketers with no technical background will be able to use them.

By then, agents will be able to help across the whole process: planning campaigns, producing assets, localization, testing messaging, analyzing performance. On the other side, customers' AI assistants will comparison-shop for them, filter out marketing spin, and hand them direct recommendations.

And so a new premium emerges: trust.

When AI is going to summarize your brand for customers, compare your offering, and decide whether to recommend you, vague, evasive messaging won't survive. Content that games the algorithm will matter less and less; content that genuinely helps customers decide will be worth more and more.

Here's what's interesting: the machines are pushing marketing back to its roots. Processes get faster and more automated, but what separates the winners is the quality of the ideas, the clarity of the story, and the strength of human-to-human connection.

In Julia White's words: we can finally get back to our craft.

AI isn't here to replace marketers. It's handing you back time, insight, and helpers — so you can do the very thing that made you want to enter this field in the first place.

In a world reshaped by AI, the best marketing may well be more human than it has ever been.