1. When the label doesn’t fit the work
When I joined TOTVS, no one sat me down and asked: are you a generalist or a specialist?
Honestly, I don’t think anyone had time for that question. There was too much to build.
In those early months, the work was almost entirely about creating structure where there was none. Explaining to stakeholders what a PMM actually does and why it matters. Building processes from scratch. Figuring out, week by week, what the role even looked like inside that company. You don’t get to specialize when the floor hasn’t been laid yet. You just move.
Then things shifted. Not dramatically, not overnight, but gradually, the role matured. The processes I’d built started running on their own. People knew what to expect from the PMM function. And with that stability came a different kind of demand: go deeper. Dig into the data. Sharpen the market intelligence. Refine the indicators that actually mattered.
I didn’t make a conscious decision to change modes. The context changed, and I changed with it.
It was only later, honestly, while writing this article, that I stopped and looked back at that path. And what I saw surprised me a little.
At no point had I chosen between being a generalist or a specialist. I had just been responding to what each moment required. The question the market keeps asking, which one are you, had never really described what the work was asking of me.
At no point had I chosen between being a generalist or a specialist. I had just been responding to what each moment required.
I don’t think that’s unique to my story. I think it’s a pattern. And I think it means we’ve been asking the wrong question for a long time.
2. When the market doesn’t know what it wants
That clarity didn’t come right away, though. To be honest, the transition itself was confusing in ways I didn’t fully expect.
Before TOTVS, I moved from Growth Marketing into Product Marketing. And I came in with a pretty clear idea of what kind of PMM I wanted to be: someone who owned launches, who lived and breathed GTM. That felt like where I could add the most value. That was the plan.
The work had other ideas.
To actually drive impact, I had to move across teams constantly. Build processes that didn’t exist. Align stakeholders who had very different views on what the PMM role was even supposed to do. The specialization I had pictured assumed a structure that simply wasn’t there yet. So breadth wasn’t something I chose. It was something the situation required.
But here’s what made it harder: the pressure to have a clear answer about which one I was kept coming from outside. Not from the work itself.
This question follows PMMs everywhere. Forums, communities, career conversations, mentorship calls. And it shows up in hiring too; sometimes, you’re asked directly: Do you see yourself as more of a generalist or a specialist? As if the answer unlocked something essential about who you are as a professional. As if there were a right answer.
When you look at senior PMM job descriptions, the same disconnect shows up more structurally. They stack expectations that belong to very different kinds of roles, as if they’re all equally urgent and equally achievable by one person.
Strategic thinking and daily operational execution. Deep customer empathy and engineering proximity. Strong analytical skills and storytelling ability. PLG experience and enterprise experience. The confidence to make calls and the patience to build consensus.
Each one of those, on its own, makes sense. Together, they don’t describe a person. They describe a department. Or the sum of every PMM a company has ever hired, wished it had hired, or imagines it might need someday. It’s a collection of hopes, not a definition of a role.
And there are real reasons this happens. Product Marketing is still relatively new in many companies, especially in Brazil. Leadership changes. New managers arrive with different mental models of what the PMM should be doing. In some cases, the organization genuinely hasn’t figured out where the PMM creates the most value. And that uncertainty ends up inside the job description.
The result is predictable. Professionals try to fit a profile that doesn’t actually exist in a coherent form. And when they struggle to fit, they assume something is wrong with them. Most of the time, the problem isn’t personal. It’s structural.
The pressure to choose a label doesn’t come from the work. It comes from outside, from the market, from job descriptions, from hiring processes.
The market knows what outcome it wants: someone who connects product and market and drives measurable results. What it hasn’t figured out is how that outcome actually gets built day to day, or what it really demands from the person in that seat.
Until that clarity exists, job descriptions will keep reading like wish lists. And the question, generalist or specialist, will keep getting asked as if it has a simple answer, when in reality it’s pointing at something much more structural.
And if the market was already struggling to define the PMM role before AI, that tension has only gotten sharper since.
3. Specialization doesn’t disappear. It migrates.
For a long time, being a specialist meant something very specific. You knew things other people didn’t. You had access to information that took time to find, frameworks that took years to develop, and pattern recognition that only came from doing the same type of work over and over. That depth was valuable precisely because it was hard to replicate.
AI changes that equation. But probably not in the way most people assume.
The most common reaction I hear is something like: if AI can do the technical work, specialization loses its value. If a tool can generate competitive analysis, produce message variations, and consolidate market data in minutes, what exactly does the specialist bring that justifies the investment?
It’s a fair question. And I think the answer is more interesting than most people expect.
AI executes well within the boundaries of what it’s been given. It finds patterns, optimizes for coherence, and synthesizes what’s available. What it can’t do is read context. And in PMM work, context is often everything.
Let me give you a concrete example. When I was building the positioning for a new solution, AI gave me technically solid message variations. Coherent, well-structured, on-brand.
But getting to a version that actually fit the research we’d done, and that would land with the real target audience, required a kind of refinement that no prompt could shortcut.
It required knowing the market, knowing the product, and knowing where the company was at that specific moment. The tool moved things faster. The judgment about what actually worked was still mine.
Same thing with competitive intelligence. AI can consolidate, compare, and organize.
But figuring out what a competitor’s move actually means, whether it’s a response to churn pressure, a push into a new segment, or a reaction to something we did, that requires hypotheses built from market knowledge that goes well beyond whatever data is available in the moment.
And sometimes the stakes are higher than just getting the message right.
Not long ago, we saw a drop in our absolute client numbers in a sector ranking. Taken at face value, that looks like a red flag. AI, fed that data point, would almost certainly frame it that way.
But context told a completely different story. We weren’t losing ground. We were dominant. The numbers just didn’t capture that on their own. Building a narrative strong enough to communicate our actual position, with precision and credibility, wasn’t an execution problem. It was a judgment call.
That’s the shift I’ve come to believe in: specialization doesn’t disappear with AI. It migrates.
Specialization doesn’t disappear with AI. It migrates. It moves away from execution and toward judgment.
It moves away from execution and toward judgment. The value of the specialist stops being about what they can produce and starts being about how well they can decide, within a specific context, what actually matters.
Less experienced PMMs can now produce more output, faster. But output and quality aren’t the same thing. Evaluating whether what AI generated is actually good, actually relevant, actually right for this moment, that requires exactly the kind of accumulated judgment that takes time to develop.
Senior professionals use AI differently. Not just to go faster, but to think better. The tool extends their reach. The responsibility for the call stays with them.
What’s genuinely being replaced is predictable execution: routine market monitoring, content at scale, data consolidation, workflow automation. These things are becoming table stakes. Anyone with the same tools can do them.
What can’t be replaced is the ability to decide what should be done, why, and when, based on context that no tool can fully capture.
The more automatable a task, the less it differentiates you over time. The more it depends on context, relationships, and situational judgment, the more it defines your value.
4. What AI can’t do: Navigate people and contexts
If specialization migrates toward judgment, then the generalist role needs a similar rethink, because being a generalist gets a bad reputation it doesn’t always deserve. It gets associated with knowing a little about everything and not enough about anything. Surface level. Scattered. Hard to pin down.
That’s not what I’ve experienced. And it’s not what I’ve seen in the PMMs who create the most cross-functional impact.
Real breadth isn’t about covering everything. It’s about having enough range to move between teams, translate different languages, and unblock problems that no single function can solve alone. It’s a specific kind of value. And it shows up most clearly in the situations that job descriptions rarely mention.
Here’s what that looked like for me across four different areas.
Technology
When I joined TOTVS, customers weren’t receiving product communications with any real focus on positioning, engagement, or demand generation. There were no structured channels. No cohesive narrative about what the solutions actually did or why they mattered.
To change that, I had to build communication channels directly inside the product, working alongside the engineering team. Pop-ups, home page content, notifications, formats for text, image, and video. It sounds straightforward in retrospect. It wasn’t.
It involved a lot of people with different priorities. It required convincing stakeholders who weren’t immediately sold on the idea. It meant bringing market context into conversations that had mostly been technical ones. And it only moved forward because I could operate on both sides, product and marketing, without losing either thread. That wasn’t a specialist skill. It was a repertoire built across different experiences.
Sales
The sales team needed more. More access to what the vertical actually offered, more concrete materials to use in conversations with prospects and existing customers, and more support across the full cycle.
Building an enablement model that actually worked meant going back and forth between sales managers and the product team constantly. Because the materials depended on inputs from both sides, and those two groups didn’t always have the same mental model of what was needed.
A content specialist could have written great materials. A sales specialist could have identified the gaps clearly. But the work of connecting those two perspectives, making sure what was being created was what the team would actually use, that was the PMM’s job. And it required range, not depth in one direction.
Marketing
My background in growth changed how I related to the marketing team. In a useful way, I think.
The PMM needs to understand which themes and levers are going to generate real opportunities. The marketing team needs product inputs they can actually build on. When that exchange works well, both functions amplify each other. When it doesn’t, they run in parallel, and neither reaches its potential.
Making it work required speaking marketing’s language well enough to be useful, and understanding product well enough to bring the right inputs at the right time. Neither of those things alone would have been sufficient.
Product
Influencing the roadmap is one of the most valuable things a PMM can do. It’s also one of the hardest to do credibly if you’re seen as someone who only brings requests.
My background in technology and IT project management helped me more than I expected. It gave me a foundation for more technical conversations with the product team. I could understand the differentiators of the solutions faster and bring customer insights in a form that was structured enough to actually shape decisions.
Without that foundation, the PMM stays at the edges of product conversations. With it, you become part of the process.
What connects these four areas is a pattern that I think gets underappreciated. In none of these situations was the problem contained within a single team. And in none of them would the solution have come from someone who could only operate within one function.
AI was present across all of it, helping with ideas, refinement, automation, and content creation. But AI didn’t notice that the problem existed in the first place. It didn’t figure out who needed to be convinced, or how. It didn’t adjust the message in real time based on how a conversation was going. It didn’t keep the initiative alive when things stalled.
AI did not identify that the problem existed, did not map who needed to be convinced, and did not calibrate the message for each stakeholder. That is the layer the generalist occupies.
That’s the layer the generalist occupies. And so far, it doesn’t have an automatable substitute.
That said, breadth has a real limit. Without a solid foundation somewhere, the generalist risks becoming a connector who can’t deliver. Someone who facilitates but doesn’t resolve. Who participates but doesn’t sustain. Breadth creates value when it’s anchored in something real, a base that gives you credibility in the conversations that actually matter.
Specialization moving toward judgment. Breadth anchored in repertoire. Two profiles that look like opposites when you see them separately. Describe the same professional at different moments when you see them together.
5. The context-driven PMM
So if the generalist/specialist frame doesn’t really hold up in practice, what replaces it?
I’ve been thinking about this a lot, and the framing that makes the most sense to me is this: the PMM of 2026 is neither a generalist nor a specialist. They are a context-driven professional.
I know that might sound like a way of avoiding the question. It’s not. It’s actually a more demanding standard than picking a side.
A generalist chooses breadth. A specialist chooses depth. A context-driven PMM doesn’t choose. They read the moment and adjust how they operate based on what the problem actually requires.
A generalist chooses breadth. A specialist chooses depth. A context-driven PMM doesn’t choose. They read the moment.
And here’s what I’ve noticed: that alternation rarely happens between projects. It happens inside the same project. Sometimes in the same week.
Let me give you a recent example.
Not long ago, the sales team at TOTVS was expanded. New managers, new salespeople, more stakeholders, more complexity. At the same time, new solutions were being added to the portfolio, which meant new things to position, new messages to develop, new content to build.
Two completely different demands landed at the same time.
The first one needed breadth. Structuring new channels, unblocking initiatives that had dependencies across multiple teams, aligning people with different expectations about what we were building and why. If I had tried to go deep on any one thing, the broader coordination would have fallen apart.
The second one needed depth. Understanding the new solutions well enough to position them accurately, adapting messaging for the right audience, making sure what we were communicating was precise enough to actually move someone toward a decision. Generic positioning doesn’t convert. It just creates noise.
I couldn’t choose between the two. Both were real, both were happening, and they required different things from me at different moments within the same context.
AI helped with both. Structuring channel content, refining messages, speeding up delivery. But the decision about when to zoom out and align versus when to zoom in and sharpen, that didn’t come from the tool. It came from reading what the situation needed.
And that, I think, is the differentiator that actually matters for modern PMMs. Not a fixed skill set. Not a label. The ability to recognize what the moment requires and shift accordingly, with enough self-awareness to know which mode you’re in and why.
High ambiguity environments, new structures, new stakeholders, things still being figured out, tend to call for breadth. Create direction. Connect the fronts that aren’t talking to each other yet. Give shape to what’s still scattered.
More defined environments, known product, mapped market, aligned team, tend to call for depth. Refine. Optimize. Drive precision where precision is what moves the needle.
The risk of being purely a generalist is losing focus when the context is already asking for precision. The risk of being purely a specialist is losing context when the environment shifts and the specialty stops being enough.
The context-driven PMM doesn’t escape those risks. But they see them coming. And they manage them actively rather than getting caught by them.
Generalism and specialization stop being identities. They become modes. What defines the professional isn’t which one they’ve committed to. It’s the awareness of when to activate each one, and the honesty to recognize when you’re in the wrong mode for the moment you’re in.
To make this a bit more concrete, here’s how each mode tends to play out in practice:
|
When to go broad |
When to go deep |
|
|---|---|---|
|
Situation |
High ambiguity, new contexts, structures being built |
Defined environment, mature processes, recognized role |
|
Objective |
Create direction and structure where there is none |
Drive efficiency, precision, and direct results |
|
PMM role |
Connect teams, align stakeholders, unblock initiatives |
Refine positioning, optimize execution, deepen analysis |
|
Key decision |
What needs to be structured, and who needs to be aligned? |
What needs to be refined to generate more impact? |
|
AI usage |
Scale coverage, accelerate ideation, map contexts |
Increase precision, refine messaging, and deepen competitive analysis |
|
Main risk |
Losing focus and generating breadth without concrete delivery |
Losing context and optimizing something no longer relevant |
|
Critical skill |
Organizational reading and stakeholder navigation |
Analytical judgment and market context mastery |
|
Expected outcome |
Initiatives moving forward, teams aligned, structure working |
Sharp positioning, strong narrative, impact on revenue, and retention |
6. Voices from the field: The same conclusion, different contexts
I want to be clear about something: the argument I’ve been making here isn’t just my own perspective shaped by my specific experience at one company in one market.
Other PMMs navigating very different contexts have been landing in similar places. Not because they read the same frameworks or followed the same career path, but because the actual work of Product Marketing keeps pushing people toward the same realization.
Two people whose thinking I respect a lot on this are Ana Paula Lafuente and Mat Silva. Both are PMA Ambassadors. Both have built careers that span strategy, execution, mentorship, and leadership. And both have thought seriously about where the generalist/specialist debate actually leads.
Ana Paula Lafuente
Ana Paula leads Product Marketing at Afya and mentors PMMs on career development. When I think about people who understand the real texture of this role, she’s one of the first names that comes to mind.
Her take on AI and the breadth/depth tension is worth sitting with:
“AI expands execution capacity and lowers barriers, allowing PMMs to navigate more easily across different disciplines. At the same time, it can also become a new specialization vertical, which I highly recommend. That is why I find this generalist vs. specialist discussion very binary.
“In the end, the PMM who stands out is the one who can orchestrate breadth and depth, connecting strategy and execution. AI comes in as an enabler, reducing operational effort and amplifying impact by optimizing time, raising the value of this professional even further.”
What I find most useful in her framing is the word orchestrate. It’s not about having both breadth and depth at maximum levels simultaneously. It’s about knowing how to bring them together in service of something larger. That takes judgment. And it takes experience. AI can support the execution, but the orchestration is still a human job.
Mat Silva
Mat works as Head of GTM & Product Lifecycle at Banco do Brasil, has won the Rising Star PMM Award, and is active as a speaker, professor, and mentor. He comes at this from a different angle, and I think his perspective adds something the conversation often misses.
“In my view, I would not put the PMM in the box of generalist or specialist. By nature, the PMM ends up being a generalist, and AI will help optimize processes and accelerate development pipelines.
“In fact, if a PMM wants to be more valued, they need to move beyond the positioning and messaging axis and focus on the real business pains that will put money on the table, looking at revenue and retention outcomes.”
That last part is important. The generalist/specialist debate can become a distraction if it keeps PMMs focused on their own profile rather than on the business outcomes they’re supposed to drive.
Mat is essentially saying: stop optimizing for the label, start optimizing for impact. Revenue. Retention. Things that actually show up on someone’s radar beyond the marketing function.
Three PMA Ambassadors, three different contexts. Health tech, financial services, retail ERP. The same underlying conclusion: the label doesn’t define the professional. What defines them is how they read context, apply judgment, and connect their work to outcomes that matter to the business.
That’s not a coincidence. I think it’s what the work actually teaches you, if you pay attention long enough.
7. How to develop this capability in practice
Here’s something I’ll admit that I think is worth being honest about.
I did not develop the ability to alternate between breadth and depth through some structured plan. There was no framework I followed, no deliberate curriculum I designed for myself. I built it by doing the work, making mistakes, and occasionally stopping to reflect on what had actually happened.
In fact, the clearest moment of understanding came while writing this article. Going back through my own trajectory, revisiting decisions I had made without fully thinking through why, and identifying patterns that were invisible to me at the time. It was only in that retrospective that the alternation became legible.
I think that’s more common than people admit. And I think it has a practical implication: this capability doesn’t develop automatically with experience. It develops when you make the invisible visible. When you stop and actually look at how you’ve been operating, not just what you’ve been delivering.
So that’s where I’d suggest starting.
Start with a retrospective diagnosis
Before thinking about what to do differently going forward, it’s worth understanding what you’ve actually been doing until now.
Not as a performance review. More like an honest audit of your own patterns.
Where have you been creating structure where none existed? Where have you been refining and optimizing what was already in place? When did each of those modes show up, and what was the context that called for it?
A few questions that I’ve found useful for this kind of reflection: In the last three months, have you spent more time building structure or improving what already exists? When your most important deliverables came together, was it because you connected the right people, or because you went deep enough on the right problem? When something stalled, was the blocker a lack of alignment or a lack of precision?
There’s no right answer to any of those. The point is to see the pattern, not to grade yourself against it.
Develop through deliberate exposure
Once you have a clearer picture of where you naturally gravitate, the development work becomes more intentional.
If you tend toward depth, look for projects that force you into broader coordination. Complex launches where multiple teams have to move together. Positioning changes that require buy-in from people with very different agendas. Initiatives where the hard part isn’t the output, it’s getting everyone aligned around the same direction.
If you tend toward breadth, push yourself into situations that demand real analytical rigor. Building a competitive analysis that goes beyond surface comparisons. Developing a data-driven narrative that has to hold up under scrutiny. Refining messaging for a segment you don’t know as well as you should.
But the type of project is only part of it. What matters just as much is how you show up inside the project. Before you start anything significant, try adding one question to your usual preparation: what mode does this context actually need from me right now? It sounds almost too simple. But asking it consistently changes how you prioritize, how you engage, and how you make decisions along the way.
Use AI as a development accelerator, not just an execution tool
This one I feel strongly about, because I think most PMMs are leaving something significant on the table.
AI changed how I work in the obvious ways. Tasks that used to take days take hours. Research that required extensive manual effort becomes faster and more comprehensive. Creating presentations, reports, and messaging drafts is quicker in ways that free up real time for higher-order thinking.
But the change that’s mattered more to me is less obvious. AI has expanded the range of what I explore. It exposes me to angles I wouldn’t have considered on my own, frameworks I wouldn’t have reached for, ways of structuring a problem I hadn’t seen before. Each well-conducted interaction teaches me something beyond the immediate task.
The PMM who uses AI only to deliver faster captures a fraction of its potential.
The PMM who uses it to think differently, to stress-test their own assumptions, to explore the edges of a problem before committing to a direction, that professional is developing something that compounds over time. Not just speed. Judgment.
And judgment, as I’ve tried to argue throughout this article, is exactly what the context-driven PMM runs on.
Over time, reading context does get more intuitive. You start to recognize the signals faster. The adjustment becomes less deliberate and more automatic. But that fluency takes time to build, and it doesn’t build itself. Especially when you’re in a new environment, a new company, a new market, the intentionality has to come back. You have to actively choose to read the situation before you react to it.
That’s the practice. Not a one-time exercise, but an ongoing discipline of paying attention to how you’re operating, not just what you’re producing.
8. The question that needs to change
I want to close with something that took me a while to see clearly.
The generalist or specialist question is not a bad question. It’s just aimed at the wrong thing.
When PMMs ask it, they’re usually trying to figure out how to grow, how to be more valuable, how to build a career that holds up over time. Those are the right instincts. The problem is that the question points toward identity, toward choosing a label and committing to it, when the thing that actually drives career growth is something different.
It’s how you develop your capacity to respond to context.
Most career development plans I’ve seen, including ones I’ve built for myself, are organized around skills. I need to get better at data. I need to sharpen my GTM instincts. I need to become a stronger storyteller. Each of those goals is legitimate. But they all share the same underlying assumption: that the differentiator comes from accumulating specific competencies in a more or less fixed direction.
If the argument I’ve been making here holds up, that assumption needs to be questioned.
The modern PMM’s differentiator is not what they know. It is the ability to read context and adjust how they operate as the moment demands.
And that capability doesn’t grow the same way skills do. It grows through exposure to different kinds of problems, different organizational environments, different stages of company maturity, and through the discipline of reflecting on what each of those experiences actually required from you.
Which means the central question of development changes, too. Instead of what skill do I need to build, the more useful question becomes: what contexts do I need to be exposed to, so that my capacity to read situations and adapt keeps expanding?
That shift is subtle. But it changes almost everything about how you pursue growth…
- Which projects do you seek out?
- Which opportunities do you say yes to?
- How do you evaluate whether a new role or a new challenge is actually going to make you better, or just keep you comfortable inside your existing mode?
AI fits into this differently than most people think about it. It’s not just a productivity tool. Used well, it’s a way to develop judgment faster than experience alone would allow.
Every time you use it to explore a problem from an angle you wouldn’t have taken on your own, to pressure-test an argument, to stress-test a positioning before it goes live, you’re building something that compounds.
In the end, the evolution of a PMM career stops being about finding the right label and starts being about expanding your range of response.
The most important question is no longer who you are.
It becomes: for the context in front of you right now, what is the best way to operate, and how do you use everything available to you to generate the most impact you can?
