Saying AI is moving at a fast pace is like saying water is wet. We’re all more than aware of that by now. 

But from the perspective of someone inside a company that deals in large part with AI tools — both for completing our own work and within our product — AI really has sped up the pace of everything we do. We spend more time using and testing AI tools than we do considering the overall impact AI is having on our jobs, our creativity, our critical thinking, and even our employability. 

Which is why it’s refreshing and important to stop and assess how AI is changing jobs in ways we may not even realize. 

AI and the future of work

Five of our Wrike sales team members recently put on a panel webinar that was a departure from our typical webinars. We usually focus on demoing a new product or outlining best practices for Gantt charts. For this webinar, we decided to slow down and offer a thoughtful discussion about how AI is changing their jobs and the people skills AI can’t replace.

We decided to roll the insights our sales team shared on the webinar into an eBook, because it really was a discussion worth diving into more deeply. The eBook is designed for a range of audiences, from teams interested in how Wrike is navigating the onslaught and onboarding of AI tools, to those considering applying for jobs at Wrike. 

The eBook is free to download, and I’d invite you to read through it. I’m really interested in this as a discussion, so let’s start the conversation here, with a glimpse into the shifts our team recognized throughout their webinar conversation. 

Keep in mind, these are just the entry point: the eBook gets into the parts that don’t compress well, like anecdotes from our team that elucidate the shifts they’re seeing, and the full checklist you can actually use day to day.

AI adoption is now the baseline

The numbers back this up: the vast majority of sales teams have already implemented AI or are actively experimenting with it, and AI now touches the start of nearly every research workflow sellers run. AI isn’t going to provide you a leading edge anymore — AI knowledge and expertise is assumed and expected.

And that means the conversation about soft skills in the age of AI has changed. When everyone has access to the same tools, the tools stop being the differentiator. What separates people now is what they do in the space those tools open up. People gathered up an AI tech stack fast, but not all of them developed the people skills to leverage in the time that tech stack freed up. 

The question now is: “AI already arrived, how are you going to use it to get even better at your job?”

Soft skills are the most durable skills

The term soft skills was probably always a mislabel. It implies those skills are less important than those that deal with the “real work.” So we’re proposing a label change to “core skills,” to describe the skills AI can’t replace. 

AI can tell you almost everything about a client from the data: history, goals, adoption patterns, risk signals. But AI can’t sit in a live conversation and sense that something is off — that the person across the table is nervous about internal buy-in, or that there’s an unspoken objection sitting in the room. Only a person can interpret a face or read between the lines or hear what wasn’t said out loud.

That’s arguably the hardest skill on the list, and it’s because AI and soft skills occupy such different territory that this one keeps its value no matter how sophisticated the tools get.

Prompting and editing are real disciplines

If you want one piece of advice for how to use AI effectively at work, stop treating AI like a search engine.

Most people’s first instinct is to ask a single question and expect a finished answer, but that’s not how you get useful output. Treat AI more like a highly competent assistant, someone smart who’s never met your client and only has the notes from your last three meetings.

A simple four-step brief solves most of this:

1. Context: Give the AI tool everything it can’t already know: account history, past conversations, the internal politics (without any incriminating details!).
2. Guardrails: Be specific about tone, length, what to avoid, and what has to be true.
3. A standard: Decide what a good answer actually looks like before you ask for one.
4. The edit: Every single output needs you to rigorously review and to strip the generic language out and put your own voice back in. This step is non-negotiable.

The people who get the most value out of AI treat prompting and editing as two halves of the same discipline, one equally important as the other. 

Judgment beats tool volume

There’s a phase almost everyone goes through and that’s trying every AI tool available because you’re worried about falling behind. This is a good instinct; in fact, it’s where most of us have started. But you’re not done yet. 

Once you figure out what tools exist, it’s time to figure out what tools will make you not just faster at your job, but fundamentally better. 

That shift, from tech stack to judgment calls, is really what critical thinking in the age of AI comes down to. AI generates enormous volumes of content, and much of it is inaccurate, not relevant, or untrue. Hallucinations are still a real, live problem. So judgment ends up doing two jobs at once: verifying that what you’re looking at is actually correct, and deciding whether it adds value. Does it genuinely improve the work? Or does it just add more words to sift through?

If you’re still in the trying-everything phase, your next move is to stop, pick a few tools that genuinely work for you, and double down.

Roles are shifting, not disappearing

A lot of the conversation about how AI is changing the job market gets stuck at the extremes — either AI replaces everyone, or nothing really changes. We’re not seeing either of those right now. 

What’s actually happening looks more like a redistribution. The backend of a lot of customer-facing work — meeting notes, data analysis, admin — has become dramatically less human. Meetings get recorded and summarized automatically. Notes take themselves. 

But the backend was the grunt work, and I don’t know many sales folks who are sad to get some help with that side of things. The frontend was always far more interesting: people-facing, strategizing, and improvising in the moment.

So what happens to the frontend when the backend runs itself with AI tools? So far, the answer, according to our panelists, looks like more time to actually think before a call, deeper conversations instead of basic ones, and more room for empathy.

The increase in AI tools has led to real shift in what jobs actually consist of day to day.

From information provider to strategic advisor

For years, customer-facing teams have tried to move from being support agents who answer questions to being strategic advisors who shape decisions. AI has pushed that shift further than almost anything before it for a simple reason: it removes the admin tasks that kept the role tactical.

A technical question that used to require a person digging through documentation now often just gets answered directly by an AI tool connected to the company’s knowledge base. Call prep that used to mean sifting through reports and product updates now takes a fraction of the time using AI. The time that gets freed up increasingly goes into the actual strategic conversation with customers.

There’s an interesting side effect here: clients are showing up more informed as well. They have access to a lot of the same tools, speeding up prep work on both sides of the table. That levels the room and produces a better conversation. This is a great example of what people mean when they talk about AI and the future of work.

AI doesn’t generate creativity

Picture a scenario where you know exactly what you want to communicate to a client, but you can’t quite find the right phrasing or the right metaphor. You feed an AI tool your half-formed thoughts and ask for options. It spits out several options and you scan until you find the one that was just on the tip of your tongue. When you use it with a client, did AI just replace your creativity?

Not really. The insight of knowing what this conversation actually needed was still on you. The judgment was still yours as well, as you decided which of the AI-generated options would actually land with this customer. AI supplied vocabulary for something you already knew you needed. The conclusion shouldn’t be that AI is creative or even that it kills a person’s creativity. It’s a tool for articulating thoughts that already have to be good before you ever open the prompt window.

Small, consistent learning habits beat sporadic AI sprints

If you’re waiting for a big block of time to finally sit down and learn AI properly, stop. The most sustainable approach to how to develop AI skills isn’t a weekend deep-dive into a particular tool, but rather a much more incremental solution than that.

Here are a few things that actually work without requiring a full schedule overhaul:

• Block 30–60 minutes a week. Small enough to remain even through a bad week, large enough to actually test something new.
• Passive learning counts. A podcast on a walk, a newsletter you scroll while you drink your coffee, following a few people on social media who post about AI consistently. Zero extra hours required.
• Ask a colleague what they’re actually using. More useful than any single tool or article, set up a quick coffee chat with your team to discuss what they’re using this week.
• Share what you find immediately, including the things that didn’t work. Nobody has AI fully figured out yet, so there’s no reason to gatekeep.

The task of learning everything about AI is both daunting and unrealistic. Instead, set small routines and let them compound into manageable learning.

Why Wrike skipped prescriptive AI use cases

When AI started showing up in our own teams, we set guardrails instead of a rulebook.

Our legal and security teams defined the essentials upfront: which tools were approved, how data could be handled, and where human review stayed non-negotiable. Those boundaries did the heavy lifting. Inside them, we didn’t dictate how employees should use AI.

We didn’t offer prescribed list of use cases or mandated workflow. That meant teams had the freedom to experiment and discover what worked for them.

That approach reflects how we think about AI across Wrike: governance earns autonomy. When the boundaries are clear and the controls are in place, people trust the system enough to delegate real work to AI, and that trust is what turns experimentation into results.

Most organizations optimize for speed or productivity and stop there. At Wrike, what we’ve seen instead is people building their own standards without being told to: personal editing rules, value tests, and an insistence that a person remains the final judge of anything AI produces.

At Wrike, we’ve skipped the prescribed rollout plan in lieu of real conversations about AI. 

A core skill stack for the AI era

When we asked our own team to distill what actually matters right now, it came down to four things. Notably, none of them are a specific tool, platform, or model, because a year from now, the tools will have changed. 

1. Critical thinking and judgment: Verify what’s correct, and ask whether it’s actually adding value or just adding fluff.
2. Adaptability: Stay comfortable learning new tools and pivoting workflows, without chasing every new thing that shows up.
3. Empathy and reading between the lines: Understand what wasn’t said in the room with customers. This is the skill AI is furthest from touching, which makes it the most durable investment on this list.
4. Proving ROI: Demonstrate real return, since budgets are tighter and tools get reevaluated more often than they used to.

In the end, this is what people skills in the age of AI actually look like in practice: a clear sense of the human elements that are elevated as AI takes over automation and admin tasks.

If you found this post useful on its own, the eBook goes even deeper and it’s free to download.