🚀 ✅ How to Roll Out Your First AI Studio Use Case to Your Team — Part 2: Handling Change Management

[Ethan_DeWaal]

Hi again, Ethan DeWaal from the AI Studio team at Asana.

In Part 1 of this series, we focused on making sure your workflow actually works: testing edge cases, getting SME validation, and making sure your automation holds up in the real world.

So now what? You’ve tested your AI workflow. It’s handling edge cases like a champ, your SMEs gave it the thumbs up, and you’re ready to unleash this automation on the world. But here’s the plot twist nobody talks about: even the most brilliantly designed workflow can die a quiet death in the “we’ll try it later” graveyard.

The hard truth? Your biggest challenge isn’t technical—it’s human. Your teammates are going to look at your beautiful automation and think: “I don’t trust it,” “It’s probably not reliable,” “I have no idea how this thing works,” or my personal favorite, “It’s faster if I just do it myself” (spoiler: it’s not).

This guide isn’t about change management theory. It’s about the actual conversations, messages, and moves that get people to trust and use AI workflows. Think of it as your playbook for turning skeptics into champions, one small win at a time.

Step 1: Pick Your Battle (Start Small or Go Home)

The Biggest Mistake People Make

Rolling out to everyone at once. “Hey team, starting Monday, all requests go through the AI workflow!”

This is how rebellions start.

The Pilot Approach That Actually Works

Find your friendly early adopters—you know who they are. They’re the ones who get excited about new features, who actually read your Slack messages, who’ve been complaining the loudest about the manual process you’re fixing.

Start with 3-5 people. That’s it.

Here’s your recruitment message that actually works:

“Hey [Name], you know how you mentioned spending 30 minutes yesterday just figuring out who should handle that client request? I’ve built something that might help. Want to be one of my guinea pigs for a week? Fair warning: it might have hiccups, but I promise I’ll handle any cleanup personally. Plus, you get to influence how this thing actually works before everyone else uses it.”

Notice what this does:

• References their actual pain point
• Sets realistic expectations
• Makes them feel special (pilot user, not test subject)
• Promises support

Step 2: Show, Don’t Tell (The Demo That Converts Skeptics)

The “Watch This” Moment

Remember those test cases from the testing guide? Time to turn them into your demo gold.

Don’t schedule a 30-minute training. Instead, catch people in the moment:

“Hey, I see you just submitted that project request. Mind if I show you something cool? Watch what happens when I run it through our new workflow… [run it live]… See how it automatically identified this needs design resources and routed it to the right team? That just saved you from three follow-up messages.”

The Comparison That Sticks

Create a simple before/after visual. Not a complex flowchart—something like:

Before AI Workflow:

1. Submit vague request (2 min)
2. Get clarification questions (wait 2 hours)
3. Answer questions (5 min)
4. Get more questions (wait another hour)
5. Finally get assigned to right person (wait until tomorrow)
6. Work actually starts (Day 2)

With AI Workflow:

1. Submit request (2 min)
2. AI asks clarifying questions immediately (answer in 3 min)
3. Auto-assigned to right person with full context (instant)
4. Work starts (same day)

Your life gets 4 hours back. Their work starts a day earlier. Everyone wins.

Step 3: Address the Fear Factor (What They’re Really Worried About)

“I Don’t Trust It”

What they really mean: “What if it makes me look bad?”

Your response: Build in safety nets and make them visible.

“Every task the workflow creates includes a note saying it was auto-generated. If anything looks weird, there’s a thumbs down button that alerts me. Plus, for the first month, I’m personally reviewing every automation to make sure it’s working right.”

“I Don’t Know How It Works”

What they really mean: “This feels like black magic and I don’t like it.”

Your response: Radical transparency.

Create a dead-simple explanation: “It reads your request → looks for keywords like ‘urgent’ or ‘design’ → checks capacity → assigns based on current workload. It’s basically following the same decision tree you would, just faster. Want to see the exact rules it follows?”

Pro tip: Offer to show them the actual workflow logic. Most won’t take you up on it, but the offer itself builds trust.

“It’s Faster to Do It Myself”

What they really mean: “I’ve learned workarounds for our broken process and change is hard.”

Your response: Focus on the awful parts they won’t miss.

“Remember last week when you had to message three different people to figure out who handles vendor contracts? Or when you spent 20 minutes reformatting that project brief because the requester used the wrong template? That’s the stuff this handles. You still do all the interesting work—this just kills the boring back-and-forth.”

Step 4: Make It Impossible to Ignore

The Redirect Strategy

Don’t shut down old channels immediately. Instead, become a friendly broken record:

• Week 1-2: “Thanks for sending this! Quick favor—can you submit this through our new workflow? [link] It’ll actually get handled faster that way. I’ll take care of it this time, but going forward that’s the best path.”
• Week 3-4: “Hey! This needs to go through the workflow now [link]. The good news is it’ll automatically get prioritized and assigned. Let me know if you hit any snags!”
• Week 5+: “Workflow link: [link]. This channel is now just for questions about the process!”

The Progress Breadcrumbs

Share wins early and often, but keep them bite-sized:

• Monday: “Quick win: The AI workflow handled 15 requests over the weekend. Zero escalations. ”
• Wednesday: “Fun fact: Average time from request to assignment dropped from 4 hours to 12 minutes this week.”
• Friday: “@channel Sarah just told me the workflow saved her from a painful requirements gathering session. Her quote: ‘It asked better questions than I would have.’”

Step 5: Measure What Matters (But Talk About What People Care About)

The Metrics That Don’t Work

Don’t lead with:

• “We’ve automated 78% of intake!”
• “Throughput increased by 45%”
• “We’re processing 3x more requests”

Your team doesn’t care about throughput. They care about their Thursday afternoon.

The Metrics That Get People Excited

Instead, talk about:

• “Zero requests sitting in limbo over 24 hours (down from 12 per week)”
• “Average back-and-forth messages per request: down from 5 to 1”
• “Time spent on ‘who should handle this?’ discussions: basically zero”
• “Requests that arrive with all the info needed: up from 30% to 85%”

The Baseline Nobody Talks About

Before you launch, spend one week documenting the current pain:

• Screenshot the confused Slack threads
• Count the “sorry, who handles this?” messages
• Track how many times someone says “I don’t have enough information”

You’ll use these as your “remember when?” moments later.

Step 6: The Care and Feeding of Your Automation

Your First Month Playbook

Week 1: The Soft Launch

• 3-5 pilot users only
• You’re basically on-call
• Fix issues immediately
• Document every question

Week 2: The Expansion

• Add 5-10 more users
• Share lessons learned from Week 1
• Start the redirect strategy

Week 3: The Momentum Build

• Open to the full team
• Daily win sharing
• Address concerns publicly

Week 4: The Victory Lap

• Share the before/after metrics
• Celebrate your champions
• Plan your next automation

The Feedback Loop That Actually Works

Don’t send surveys. Nobody fills out surveys.

Instead, drop into people’s DMs:

“Hey, I noticed you’ve used the workflow 5 times this week. Be honest—is it actually helping or just different? What’s the one thing that would make it 10x better?”

The magic phrase: “What’s the one thing…” It gives people permission to criticize while keeping feedback focused.

The Scripts You’ll Actually Use

For the Skeptical Manager

“I know you’re concerned about reliability. That’s why we’re starting with just internal requests where mistakes are easy to fix. Once we prove it works for 2 weeks with zero issues, we can discuss expanding. Fair?”

For the “Old Way Works Fine” Person

“You’re right, the old way does work. This is just about eliminating the annoying parts. You know how you always have to chase down budget approval info? The workflow now requires that upfront. You never have to ask again.”

For the Worried Over-Sharer

“I get that you like to provide context in your requests. Keep doing that! The AI actually performs better with more information. It’s the people who write ‘need help with thing’ who’ll have to adjust.”

For the Secret Workflow Hater

“Look, I know this isn’t your favorite change. What if we agree on a trial period? Use it for two weeks, and if you genuinely think it’s making your work harder, we’ll figure out an exception process. Deal?”

Your “Is This Actually Working?” Checklist

Two weeks after launch, you should be able to say yes to:

People are using it without being reminded (mostly)
The complaints have shifted from “I don’t trust it” to “Can it also do X?”
At least one person has said something nice about it unprompted
Request quality has noticeably improved
You’re not personally handling issues daily anymore
Someone asked if you can build another workflow for a different process

If you can’t check these boxes, don’t panic. Go back to Step 3 and really listen to what people are saying. The solution is usually simpler than you think.

The Real Talk

Here’s what nobody tells you about AI automation: The technology is the easy part. The hard part is convincing talented, smart people that this tool will make their work life better, not more complicated.

Your workflow might be a technical masterpiece, but its success depends on how well you handle the deeply human fears about change, control, and being replaced by robots (spoiler: you’re not replacing anyone, you’re just eliminating the parts of work that everyone secretly hates).

Start small. Celebrate tiny wins. Be annoyingly transparent. And remember—every person who goes from “I don’t trust this thing” to “How did we live without this?” becomes your biggest champion.

Your automation is ready. Your test results are solid. Now go turn some skeptics into believers.

One last thing: In three months, when everyone’s forgotten how painful the old process was and your AI workflow is just “how we do things,” take a moment to celebrate. You didn’t just build an automation—you changed how your team works. That’s the real innovation.

---

Missed the first part? Check it out here

🚀 ✅ How to Roll Out Your First AI Studio Use Case to Your Team — Part 1: QA Checking Your MVP

Want to be automatically notified when other AI Studio Tips go live?
Make sure to follow AI Studio Tips & Workflows!

---

[lpb]

Thanks, @Ethan_DeWaal!

These are practical, actionable, and tremendously comprehensive tactics covering all the bases.

It sounds like you’ve done this before (Kidding; I’ve been appreciating your published work here and elsewhere since the start.)

Thanks for this and for Part 1, which was great too,

Larry

[Sarah_Yeung]

Just want to drop a massive for thanks for bite sized step by step for introducing change to a team. As a newly joined marketer PM who’s spearheading Asana after a flopped roll out company wide almost 2 years ago now, retroactively trying to achieve buy in and suggest change to redundant processes has been by and far the hardest part and I’m always appreciative to read and hear of other people’s povs in people management.

So simple.. and so hard right ?