Free series: Ship It Explore turning your expertise into a product Explore it →
Communication

Why Communication Is Not a Soft Skill — Even in the Age of AI

· · 6 min read

I don’t dress up that often. I grew up a tomboy. I hated wearing dresses and makeup. But my youngest is such a girly girl, and over the past couple of years she’s inspired me to dress up and celebrate. So this past weekend, I got all gussied up for my dad’s birthday. It turned into a reminder that communication is not a soft skill, even though we keep filing it under that label.

We had a fun dinner, the kind that included kulfi made fresh at our table for dessert. (Kulfi is an Indian ice cream.) And my 73-year-old dad loved it. You’d think these little things would get boring as you get older, but he embraced the moment and thought it was pretty novel.

Poornima Vijayashanker dressed up for her father's birthday dinner.

Feasting with family can feel indulgent. But it’s what keeps us together, and it’s what we reach for during the hard times.

Here’s what stuck with me afterward. We’re quick to wave off things like this as soft. Indulgent. A nice-to-have. The dinner, the dressing up, the dessert made tableside. But they aren’t extra. They’re the glue.

And it got me thinking about another thing we’re far too quick to file under “soft”: communication.

We call it a soft skill. Then we spend nearly every waking hour doing it.

Look at your actual day. Any job, any role. You figure out what needs doing. That’s a to-do list, which is you communicating with yourself. You set expectations for when it’ll be done. You prioritize. You flag the scope creep and the thing you’re blocked on.

Then you do the work. And the second you get stuck, what happens? A meeting. More communication.

The only stretches of my day without some form of it — talking, listening, reading, writing, live or async — are when I’m sleeping, eating, or getting ready. That’s the “soft” skill. The one running underneath basically everything else.

Even AI hasn’t made it less important. If anything, it raised the bar.

Every time you prompt an agent, review what it gives back, and correct what it got wrong, you’re communicating. But being clear isn’t always enough. Sometimes it runs off and does the thing you didn’t say instead of the thing you meant. You have to spell out that you’re about to give step-by-step instructions, and then actually give them.

Here’s the difference from a coworker. A person reasons about consequences. They hear your request and think, “if I do this, that other thing breaks, so let me check first.” The model doesn’t. It has no stake in the outcome and no instinct for what happens next.

So the entire burden of precision shifts to you. That’s not less communication. That’s more.

Even if you think human communication can be deprioritized, the entire race to build better AI is a race to build a better communicator.

And here’s the part that gets me. Why are companies pouring billions into improving these models? To make them better at taking in input, interpreting it, and delivering output clearly. Better communicators.

And how do they get them there? By feeding them more human communication. The defining technology investment of our era is, at its core, a giant bet on the very skill people love to call soft.

So maybe it’s time to stop calling it that.

Let me make it concrete, because I lived it

The iPhone 18 Pro just launched. What you didn’t see is that I worked on that product right up until I left Apple this past April.

We were resource-constrained, which meant the team had to be careful with what we had. At the start and end of every day, I synced with colleagues in China. I couldn’t just hand them a list of what I’d finished and what I needed. I had to be clear on the details: what was working, what wasn’t, what to prioritize, what could wait, and what we needed to watch so it wouldn’t derail the project or, worse, cost us rework.

Translating engineering problems into next steps was where I spent the bulk of my time and energy. Sure, I used AI to stay organized and draft my weekly report. But I was the one who got pinged the moment there was confusion, or someone had to be brought up to speed.

Things move fast in iPhone land. The key was always the same: knowing who I was talking to, what they needed to know, what I could leave out so I didn’t overwhelm them, and what I needed them to do.

The framework I ran every day

Think about the last piece of work you had to redo. Was it really a skills gap, or was it a miscommunication, with a coworker or with AI?

Most rework starts there. Here’s the framework I ran every single day at Apple to keep it from happening. Before you send the update, the message, or the prompt, get clear on two things.

First, the audience

  • Who am I delivering this to?
  • What do they actually need to know?
  • What can I leave out so I don’t overwhelm them?
  • What do I need them to do?

Then, the substance

  • What’s the problem? (And say it differently for engineers, senior leadership, and customers.)
  • What’s been triaged or debugged?
  • Where are we still stuck?
  • What are the suggested next steps?
  • What’s the risk?
  • What’s the expected delivery date or rollout cadence?

Rinse and repeat. Not every update needs a deep dive or more experts pulled in. Knowing when it does is its own kind of intuition, and it’s learnable.

And it pays off directly: the clearer the communication, the faster you get to a decision, or to the next step toward solving the problem. Even deciding the problem isn’t worth solving. 😉

That framework is just one tool

Communication is really a whole toolkit: telling a story, leading with the bottom line, handling the hard question, turning jargon into an analogy someone outside your world can actually hold onto. The real skill is knowing which one the moment calls for.

And notice the thread running through all of it. Whether you’re introducing yourself, giving a status update, delivering a lightning talk, running a workshop, or just having a daily one-on-one, it comes down to the same thing: understanding the person on the receiving end. Their context. What they need. What they’ll do with it.

Here’s the hard part. We rarely get concrete feedback on how a message actually landed.

You send the update, give the talk, run the meeting, and mostly you’re left guessing whether it worked. That’s exactly why I’m teaching this workshop. It gives you the space to get your reps in, get real feedback, and see how other people do it too.

Come practice this week

This Friday, September 18, I’m teaching a free, live 90-minute workshop that’s a taste of the Confident Communicator Course. It’s capped at 20 people, and there are just a few spots left. If that’s you, grab one before they’re gone.

Save your seat for the workshop here →

I recently joined executive coach Jamie Lee on her podcast, Risky Conversations, to talk accountability, career agency in the age of AI, and why your expertise is worth more than you think. Have a listen, or watch on YouTube.

Pocket
Share on reddit
Share on LinkedIn
Bookmark this on Digg

Poornima Vijayashanker

Founding engineer at Mint.com. Senior SWE & EPM at Apple. Building communication systems for technical professionals.

The Femgineer Newsletter

Technical communication, delivered weekly.