Pause to Go Faster: Designing 'Valuable Friction' for AI at Work

Why smart teams speed up by choosing the right moments to slow down

Sep 10, 2025

Recently, I attended CEX, the Content Entrepreneur Expo, in Cleveland, Ohio. Two speakers — Robert Rose and Ann Handley — made a case that can sound “anti-AI” at first blush: in a world racing to automate and accelerate, we’re undervaluing the moments that deserve time.

Their premise is not anti-technology, however, but anti-thoughtless speed.

Rose calls for valuable friction—the purposeful pauses that lend meaning to work. Handley urges us to find slow moments—the points in a process where going slower improves outcomes.

In his new book, Valuable Friction, Rose says:

We live in a world obsessed with speed. But in our rush to optimize, automate, and move on, we often lose the very thing that makes our work meaningful: the friction.
Valuable Friction is a manifesto for doing things differently. It’s not about resisting progress—it’s about designing better resistance. Whether you’re building a team, leading a brand, or navigating your own creative chaos, this book offers a fresh lens on what happens when we pause, push back, and give things the time they deserve.

In her forthcoming book, titled ASAP: The Magic in Going Slow at Just the Right Moments, set for release in 2026, Handley explores how strategically slowing down can lead to greater success in a fast-paced world.

In our always-on, mile-a-minute, accelerated world… we’re often missing out by moving too fast. So fast, that we don’t know when we’re missing out—and what we’re missing out on.
Some things do need a slower, more thoughtful approach. But what’s more: There’s magic in knowing which moments to embrace as Slow Moments.

Applied to AI in the workplace, their shared premise is simple and radical: let machines shrink the busywork so humans can stretch the thinking.

The risk isn’t AI itself; it’s using AI to do more of the wrong thing, faster if you’ve felt a content volume spike while quality, trust, or team morale dip, you’ve met the problem.

The solution isn’t to slam the brakes. It’s about designing the brakes to build intentional resistance into the right parts of your workflows so that you can drive faster, safer, and with a clearer destination.

Below is a practical playbook for leaders who want to operationalize Rose’s and Handley’s advice across AI-enabled teams.

👉 Could you spare 2 minutes to complete our AI Technostress survey

Help us make AI adoption human-friendly. 2 minutes. 10 questions. Real impact.

Take the Survey

A Simple Operating Principle

Fast where machines excel. Slow where meaning is made.

Fast lanes: repetitive, deterministic, “good-enough” tasks (summaries, formatting, transcriptions, first-draft scaffolds, data extraction).
Slow lanes: high-stakes choices, brand voice, original insight, ethical risk, customer promises, and anything that changes behavior or reputation.

Your job is to label the lanes and install the guardrails.

12 Steps to Design Valuable Friction

1. Map the work: label Fast vs. Slow

Gather one representative workflow from marketing, sales enablement, customer support, HR, and analytics. Break each into steps. Mark every step as F (Fast) or S (Slow) based on these questions:

Does this require original judgment, ethical trade-offs, or brand voice? → S
Is it repetitive pattern-matching or data munging? → F
Would a wrong decision here carry legal, trust, or safety risk? → S

You’ll see quickly where you’re overspeeding (too much “F” in “S” territory) and where you’re under-automating.

2. Define your Slow Moments

Write a one-page policy that names the specific moments that must not be rushed. Examples:

Positioning decisions (core messaging, product claims, pricing changes)
Customer-facing claims that require sources or approvals (compliance, medical, financial, security)
Sensitive comms (crises, layoffs, policy shifts)
Original research (methodology, interpretation, conclusions)

Attach a simple rule: “If it’s on this list, it sleeps at least one night.” That alone prevents a shocking percentage of preventable mistakes.

3. Use an AI brief before you use an AI model

Valuable friction begins before prompting. Require a 5-question “anti-rush” brief for any AI-assisted output:

Who specifically is this for?
What change do we want (think/feel/do)?
What truth and proof do we have (sources, data, examples)?
What’s non-negotiable in voice/tone?
What are the known risks (accuracy, bias, compliance)?

No brief, no prompt. The model is only as thoughtful as the inputs you provide.

4. Install preflight checklists at publish points.

Embed lightweight, two-click checks in your CMS, helpdesk, or sales-asset tools:

Citations present? (source links or internal references)
Brand voice pass? (style guide keywords/phrases confirmed)
Risk tag applied? (low/medium/high with route to approver for medium+ high)
Human reviewer named? (no “anonymous approval”)

The checklist is friction that pays dividends.

5. Tier your human-in-the-loop

Create three review tiers:

Tier 1 – Ship: Low-risk, low-impact items auto-publish after basic checks.
Tier 2 – Review: Medium-risk assets get one human editor and automated fact-checking.
Tier 3 – Deliberate: High-risk work involves legal and compliance considerations, as well as an “overnight rule.”

This keeps most work fast while protecting the moments that matter.

6. Train for slow skills, not just fast tools

Your AI program should teach questioning, synthesis, and ethical reasoning alongside prompting tactics. Pair modules like:

“From output to argument: turning AI drafts into defensible POVs.”
“Bias spotting 101: how to challenge your model and yourself.”
“Source hygiene: tracing claims to evidence.”

These are the muscles of valuable friction.

7. Measure quality, not only velocity

Add metrics that reward Slow Moments:

Error rate and post-publication corrections (goal: down)
Brand coherence score (consistent voice and claims)
Time reallocated to strategy (hours saved → hours invested)
Employee technostress index (pulse checks on overwhelm and clarity)

If you only track throughput, speed will always win, even when it shouldn’t.

8. Adopt the ‘Overnight Rule’ for high-impact comms

If a message can materially affect trust, revenue, safety, or reputation, it must survive one sleep cycle. Cognitive science backs it: distance improves judgment. Build this into your SLAs so teams don’t feel punished for doing the right thing.

9. Institutionalize provenance and approval

Maintain a simple register: which models are used where, prompts/templates used, and who approved what. Add content provenance (e.g., C2PA) where possible. This is friction that creates audit trails and protects you when questions arise.

10. Make it safe to tap the brakes

Normalize a “slow flag” anyone can raise without penalty. If a writer, analyst, or PM senses risk or confusion, the work automatically moves to Tier 3. Culture is the difference between lip-service friction and lived friction.

11. Automate the trivial to buy time for the vital

Aggressively apply AI to free time for Slow Moments: meeting notes, research sweeps, outline generation, sandbox ideation, QA suggestions, localization scaffolds, spreadsheet wrangling. Put the reclaimed hours on the scoreboard and dedicate them to craft, customer interviews, and decision quality.

12. Pilot, don’t preach

Pick one workflow (e.g., product launch messaging, sales playbook updates, or customer-service macros): baseline quality and cycle time. Apply the friction steps for 30 days. Compare before/after: fewer reworks, clearer POV, better customer responses, calmer team. Publish the win internally.

A Pocket Heuristic Your Team Can Remember

Teach SLOW as a quick trigger before publishing anything important:

S—Select: Is this a named Slow Moment?
L—Listen: What do customers/stakeholders need to hear, not what we’re eager to say?
O—Observe: What’s the evidence? Are there claims without sources?
W—Wait: Would an overnight pause improve judgment?

If SLOW says “pause,” you do.

What This Looks Like in Practice

Marketing: AI drafts the long post and a carousel outline in minutes. The team spends its time on the argument, the proof, and the story arc, rather than focusing on first-pass wordmithing. One overnight pause turns a generic piece into a distinctive POV your competitors can’t copy-paste.
Customer Support: AI suggests macro updates based on ticket trends and patterns. A human reviews the top 10 changed macros weekly to ensure tone, clarity, and address edge cases. Publish fast for low-risk updates; deliberately slow for anything that alters policy or escalations.
Analytics: AI surfaces anomalies and drafts narratives. Analysts slow down to validate sources, question causality, and propose decisions. Leaders get fewer dashboards, more decisions.
HR & internal comms: AI compiles FAQs; humans shape the message with empathy. Anything sensitive follows the Overnight Rule and Tier 3 review.

The Bonus Benefit: Less Technostress, More Trust

Unbounded speed creates cognitive overload: endless drafts, constant pings, and unclear standards.

Designed friction reduces decision fatigue, clarifies what “good” looks like, and enables teams to focus—the result: higher-quality outputs, fewer reworks, safer decisions, and calmer people.

Rose’s “valuable friction” and Handley’s “Slow Moments” aren’t nostalgic pleas to put the genie back in the bottle. They’re modern operating principles for AI-enabled organizations that want to move fast with wisdom, not merely fast.

Let AI sprint. Let humans steer. And choose, on purpose, when to press the brakes.

Quick-start checklist (copy/paste to kick off)

Map one workflow and label F/S steps
Publish your Slow Moments list (one page)
Require a 5-question AI brief before prompting
Add preflight checks and three review tiers
Institute the Overnight Rule for high-impact comms
Track quality metrics (errors, coherence, reallocated time)
Run a 30-day pilot and share the results

Implement those, and you’ll feel the difference within a month: fewer “urgent” rewrites, clearer decisions, and a team that’s proud of the work again… because you designed the right resistance.

It’s not anti-AI; it’s anti-burning out employees with a hyper-intense workload.

What do you think? Leave a comment to let us know.