How I Use AI to Write Emails, Docs & Feedback
And why it finally made writing feel... not painful
TL;DR:
I used to procrastinate on writing—especially the kind that really mattered. AI didn’t just make me faster—it made me feel supported. In this piece, I share how I use AI to write emails, documents, and feedback in my day-to-day as an engineering manager, including the moment it helped me turn a looming performance review deadline into a solid, thoughtful win—for me and my team.
Performance review season is never easy—especially when you’ve got ten direct reports and a calendar packed with everything but time to sit down and write.
The first time I used AI to help with performance evaluations was in January 2024. I had ten reviews to complete, and a few of them were for team members up for promotion—so I needed to go deep. Solid examples. Thoughtful reflections. Clear rationale for advancement. Usually, I’d spend two to three full working days getting these done.
But that year, I was swamped. I didn’t have the luxury of time. What I did have was an AI assistant—and a bit of curiosity.
So, I tried something new. I opened up a blank chat and started sharing my thoughts.
I wasn’t asking AI to write for me—I was asking it to shape what I already had.
My feedback, my perspective, my arguments—I poured it all in. And we refined it together.
In the end, I wrote all ten performance reviews in about two and a half hours.
And they were good. Clear. Specific. Human.
That moment changed everything. AI stopped being a shiny tool I played with—and became a real collaborator in my work.
From Procrastination to Progress
I used to put off writing. Especially documentation, internal notes, or anything that required a bit of structure and polish. I’d tell myself I needed to be in the right frame of mind—but really, it was the weight of doing it all alone.
With AI, I don’t feel alone anymore.
It’s like having a smart, patient writing partner who’s ready whenever I am. Someone who helps me think through structure, tone, and clarity—so I can focus on what matters most.
Writing With AI: My Everyday Flow
Here’s how I typically use AI today in my workflow:
1. Start with the end in mind.
What’s the goal of this message? Who’s it for?
2. Provide strong context.
This is where the “magic” happens—clear input = clear output.
3. Collaborate and refine.
I guide the assistant like I would a co-writer. I rarely use the first draft.
4. Own the outcome.
The final version still sounds like me—because I edit it until it does.
But here’s the thing—just because I can use AI for writing doesn’t mean I always do.
When I Don’t Use AI
You might be wondering—do I use AI for all of my communication?
The short answer is no. Most of my day is filled with conversations—on Slack, in meetings, on calls, or in person. And I don’t route all of that through AI. I don’t want to.
AI is great when I need to write something thoughtful, structured, or high-stakes—like a performance review, a memo, or a cross-team email. But for day-to-day back-and-forths? Quick updates? Talking through an issue live? That’s all me.
That said, there are moments where I’ll draft a tricky Slack message or quick update in AI first—just to check tone or wording before I hit send.
The goal isn't to polish everything until it sounds perfect.
It’s to use AI where it helps me think more clearly or communicate more intentionally—not where it would make me sound less human.
It’s about balance.
AI supports my communication, but it doesn’t replace my voice. And it never should.
A Quick Note on Privacy and Sharing
Before we dive into real examples, I want to acknowledge something important.
Many people have valid concerns about using AI tools, especially regarding what happens to the content they type into a chat assistant.
Many popular AI tools—especially free or public-facing ones—do collect and retain the data you input. That data can be used to improve the model’s performance over time, which may include fine-tuning future versions, training new models, or enhancing the system’s responses.
In other words, what you type into a public AI assistant may not stay private.
Unless you're using a tool that explicitly disables data logging or offers an enterprise-grade privacy agreement, you should assume your input could be stored, analyzed, and potentially used to retrain models.
If that makes you pause, it should.
It’s worth being thoughtful about what you share and where.
For example, if I need to write any communication that contains sensitive data—like performance metrics, internal numbers, or strategic insights—I do not use public AI assistants. Instead, I use my company’s internal AI platform, which provides secure access to API-based models. These tools are configured with enterprise-grade privacy policies, which means the data stays private and is not used for model training or exposed externally. That gives me the confidence to work with real data without the fear of it leaking.
If I’m using public models, I’m much more cautious. I avoid sharing any confidential information and will often use placeholders or generalized inputs to get the structure or tone right. Then I’ll plug in the actual data manually afterward.
The key is to be intentional.
Use the right tool for the right job. When in doubt:
Sensitive or private? → Use internal tools with enterprise plans.
General or non-sensitive? → Public tools are usually fine.
Just don’t let convenience override judgment.
Real Examples: How I Use AI to Write Smarter, Faster, and Better
It's one thing to say “use AI to help you write.”
It’s another to actually see how that works in practice.
Here are three real scenarios I’ve worked through—plus the prompts I used and the results that came out of it.
1. Writing a Polished Team Email: Launching an Internal Newsletter
The scenario:
We were preparing to launch a new internal newsletter for the engineering team—a bi-weekly space for sharing updates, learnings, and moments worth celebrating. I needed to craft an announcement email that would inform, excite, and invite participation.
Prompt I used:
"Help me write a team-wide email announcing a new internal newsletter for our engineering org. The first issue will go out on [DATE]. It's a bi-weekly newsletter that will cover: tech updates, learnings, industry practices, team wins, celebrations, and more. Keep the tone warm, clear, and a little fun."
The AI-generated draft (refined):
Subject: Introducing ZALORA Tech Threads – Our New Engineering Newsletter
Hey team,
I’m excited to announce that we’re launching a new internal newsletter for our engineering org: ZALORA Tech Threads.
The first edition will hit your inbox on [DATE], and you can expect it every two weeks after that.
What’s inside?
Team updates
Engineering learnings
Industry best practices
Shoutouts and wins worth celebrating
And a few surprises
The goal is to keep us connected, inspired, and in the loop.
Got something to share? We’d love to feature it—just reach out!
—Velina
2. Writing a Meeting Invite with Context: Demo Day
Note: Feature names have been adapted for illustrative purposes.
The scenario:
We were organizing a Demo Day—a 60-minute session where engineering squads showcase new features. I needed a meeting invite that would clearly explain the purpose, include highlights of what’s being demoed, and speak to a broader audience beyond just tech.
Prompt I used:
"Help me write a Google Calendar invite description for Demo Day. It’s a 60-minute internal session where engineering squads showcase new features. This round will include: 1) Smart Cart Reminders to reduce checkout drop-offs, 2) Personalized Homepage Modules based on user behavior, and 3) Faster Refund Flow for returned items. Audience includes tech leads, designers, PMs, business stakeholders, and C-level execs. Add a friendly tone, and mention that the deck is attached."
The AI-generated draft (refined):
Join us for Demo Day – April Edition
Come take a look at what our engineering squads have been building!
This 60-minute session will showcase three key feature updates:
Smart Cart Reminders – Nudging users at the right time to reduce checkout drop-offs
Personalized Homepage Modules – Surfacing content based on user behavior and engagement
Faster Refund Flow – Making refunds smoother and speedier for our customers
Demo Day is about visibility, celebration, and collaboration—so whether you're in tech, product, or business, there's something here for you.
Deck attached. Join us, ask questions, and let’s build momentum together.
3. Writing a Strategic Memo: Increasing AI Adoption
Note: Final content is internal and proprietary. The structure and process are shared as a reference.
The scenario:
I needed to write a company-wide memo focused on increasing AI adoption—not just for the engineering team, but across business units as well. The memo needed to reflect our current state, capture key challenges, outline our goals, and propose a cohesive strategy for enabling wider adoption.
Prompt I used:
"Help me structure a strategic memo about increasing AI adoption across the company. Include sections for current state, tool usage, pain points, goals, and a proposed strategy. The strategy should include high-level initiatives to increase AI literacy, improve access to tools, and encourage adoption across teams. Keep the tone clear, future-focused, and practical."
The AI-generated structure (refined):
Memo: Accelerating AI Adoption at [Company Name]
Context & Current State
Key Pain Points
Goals for 2025
Proposed Initiatives
Next Steps & Ownership
The final memo was built through continuous refinement and deep iteration. While the structure was co-created with AI, I worked point by point to clarify the strategy, go deeper on pain points, and ensure each proposed initiative felt actionable and aligned with our goals. The version shared here is illustrative—the actual memo remains internal—but the process reflects a repeatable way to move from rough input to a solid, strategic document.
What People Get Wrong About Writing with AI
Most people think you just give AI a prompt and voilà—it spits out the perfect message.
But that’s rarely the case.
I see it at work a lot: someone uses a generic prompt, copies the first response, and sends it as-is. It might save them time, but the result feels cold, robotic, and… obvious.
Good communication needs nuance.
AI won’t get there on its own—you have to guide it.
That’s what I think most people get wrong: they rely too much on AI and skip the part where you still have to think. You still have to refine, shape, and co-create.
That’s how you get something that sounds like you—not ChatGPT.
For Beginners: How to Start Using AI to Write
If you’re curious but don’t know where to begin, here’s what I’d say:
Just start.
Pick something real and coming up in your day:
Your next email to your team
A memo or meeting summary
A process doc or strategy outline
Open a chat window and start working with the assistant.
Give it context. Ask it to help you shape your draft. Try a few edits. See how it responds.
Then—this is key—ask it to generate a reusable prompt based on what worked. Save that prompt in your own library so you can get to good results faster next time.
That’s what I’m building now—a personal set of go-to prompts that reflect how I work.
I wish I’d started sooner. And if you’re reading this, maybe now is your time.
One Last Thing
The real benefit of writing with AI isn’t just speed.
It’s the support.
It makes the blank page feel less daunting. It helps me show up for my team when I’m short on time. It brings momentum to tasks I used to dread.
AI doesn’t take my voice away—it helps me find it faster.
And honestly? That’s the most human use of technology I can think of.
Let’s Make This a Conversation
What’s one piece of communication you’ve been putting off—an email, a doc, a message—that AI could help you with today?
Hit reply and let me know. Or forward this to a friend who’s curious but stuck.
No hype. Just real people figuring it out together.
Until next time,
Velina
Version of You – Redesigning how we live and work, one update at a time.