AI was meant to bring relief. Instead, it triggered an expectation spike: leadership now assumes creative should be faster, higher quality, and more prolific—by default.

So even when production gets quicker, teams often feel worse, because what scales fastest isn’t output.

It’s noise: more stakeholders, more “quick tweaks”, more review loops, more context to chase, more surface area for rework.

That’s the creative reset: AI accelerated making… and exposed how fragile delivery really is.

Getting through it takes intervention—not more tools.

What the expectation spike looks like in the real world

It shows up as a handful of “small” shifts that quietly change everything:

• Faster becomes the baseline, not the advantage.
• Options become cheap, so the ask becomes “show me five more.”
• Polish becomes mandatory even for internal work.
• “Urgent” becomes normal, not exceptional.
• Approvals expand because more people want a say (or feel responsible for risk).
• Context becomes the work (briefs aren’t brief; the truth is spread across threads, calls, and memory).
  

None of this is about talent. It’s about the operating model not catching up.

Why it feels worse: “work about work” eats the gains

AI can help draft, resize, version, localise, remix.

But it doesn’t automatically reduce:

• stakeholder sprawl
• ambiguous briefs
• subjective feedback
• searching for the “latest final” file
• meetings that exist purely to create alignment
• “just one more thing” requests that fracture focus
  

That’s why teams can add AI and still feel underwater: time leaks into coordination.

Asana describes this as “work about work”—the admin, chasing, tool-juggling and status-wrangling that sits on top of skilled work (and can swallow a huge chunk of the week).

So the gains from faster production get consumed by a bigger coordination tax.

Noise scales faster than output.

AI can intensify work if the system stays the same

There’s a common assumption that AI reduces workload.

But research is increasingly pointing to the opposite pattern: when a tool makes a task faster, organisations often respond by expanding the scope, increasing volume, or raising the quality bar—so the work intensifies instead of disappearing.

This is why “more capability” can feel like “more pressure”.

Because the constraints moved upstream and downstream: intake, prioritisation, reviews, governance.

Budget reality makes the reset sharper, not softer

While expectations rise, many teams aren’t being given unlimited headcount or unlimited agency spend to match.

Gartner’s 2025 CMO Spend Survey press release notes marketing budgets remaining flat as a share of company revenue, with CMOs pursuing productivity gains.

So the organisation doesn’t just want “more output”.

It wants more output with less drama.

Which means the intervention has to be structural.

The new kind of noise: polished, low-signal output

AI also introduces a specific flavour of noise: content that looks finished but creates extra work for everyone else—because it’s vague, generic, or subtly wrong.

Axios summarised research on “workslop”: low-quality, AI-generated memos/emails/reports that shift the burden onto recipients to interpret, correct, or discard.

That’s not an argument against AI.

It’s an argument for where AI belongs and how outputs get governed—so you get acceleration without flooding the system.

The intervention: reduce noise first, then scale output

If you try to “AI your way out” without changing how work flows, you just hit the wall sooner.

Here’s the intervention Perpetual-style—direct, low ceremony, designed to plug in without rewriting your whole organisation.

1) Fix intake so “urgent” stops meaning “random”

If the front door is Slack + hallway asks + last-minute pings, the team will never feel calm.

Put a simple triage rule in place (one that leadership understands):

• revenue impact
• customer impact
• launch deadlines
• legal/compliance risk
• reputation risk
  

Everything else goes into the queue—not into someone’s afternoon.

2) Reduce stakeholder noise with one accountable owner

Most delivery pain is “too many hands” pain.

For each request:

• one accountable owner
• one approver who can actually approve
• one definition of done

This alone cuts rework because decisions become clearer and earlier.

3) Make reuse real: tokenise what repeats

Guidelines don’t ship work. Building blocks do.

Tokenise the repeatables:

• templates and layout systems
• components (slides, sections, UI patterns, campaign modules)
• brand rules (type, colour, spacing, hierarchy)
• approved copy modules (proof blocks, disclaimers, product phrasing)
• image style rules (composition, lighting, constraints)

This is where speed compounds: you stop reinventing and start deploying.

4) Put AI in a lane (and keep judgement human)

AI wins where the work is repeatable and bounded:

• versioning, resizing, localisation
• first drafts and variations
• tagging/organisation
• QA passes and consistency checks
  

Humans stay responsible where trust matters:

• narrative choices
• brand nuance and persuasion
• risk-heavy claims
• executive-level polish
• stakeholder alignment

AI becomes a force multiplier—not a debate generator.

A 30–60–90 reset plan for teams under pressure

Days 0–30: Stabilise

• one intake route
• visible queue + priorities
• clear ownership per request
• cut meeting creep (replace with status clarity)
• don’t bulldoze review culture yet—learn how decisions actually happen
  

Days 31–60: Simplify

• templates for your highest-volume asset types
• “definition of done” per asset class
• feedback rubric (what good feedback looks like)
• start tokenising repeat elements
  

Days 61–90: Scale

• streamline approvals based on where work actually stalls
• introduce the AI lane with guardrails + training
• build client-owned image libraries and variation kits
• standardise how the brand gets deployed across teams
  

The goal isn’t “more AI”.

The goal is predictable shipping.

Signs you’re through the reset

You don’t need fancy dashboards to know it’s working. You’ll feel it:

• fewer “urgent” interruptions
• fewer late-stage stakeholder surprises
• fewer review loops per asset
• clearer ownership and faster decisions
• less time spent searching, chasing, re-explaining
• more time for actual thinking and creative direction
• work ships without needing heroics
  

That’s the outcome leadership actually wants.

The point

AI didn’t just change tools. It changed expectations.

The teams that come out stronger won’t be the ones “using AI the most”.

They’ll be the ones that intervene to reduce noise, redesign delivery, and make speed repeatable—without burning people out.

If you want a team that can plug in quickly, cut the noise, and implement this while the work keeps moving—that’s the lane Perpetual operates in.