What happens to your marketing career when the platform you built your expertise on is replaced, or when an AI agent takes over the tasks that used to justify your role?
The usual answer is to learn another tool. That keeps your skills tied to the next platform, leaving the underlying exposure in place. A better approach comes from an unexpected place: how you design a martech stack.
A martech stack gains from change when its parts are replaceable, its data is owned, and its constraints are treated as design input. A marketing career can be built on the same three properties.
Marketers evaluate marketing technology across many dimensions. They consider whether a platform helps or hinders the marketing processes it supports, followed by cost, integration effort, and data access. Further down the list is what happens when the martech vendor is acquired, an API is restricted, or a tracking method disappears. The answers determine whether a change costs the team a bad quarter or a bad year.
AI makes these questions more relevant for careers. Over the past year, AI moved from a feature of marketing tools to systems that handle a huge share of marketing work. I’ve spent years working out what makes a marketing system gain from change instead of merely surviving it, and the more I consider the career question, the more the two look similar.
A career is exposed to change just as a stack is. The fix can come from the same skills: building portable expertise, owning your professional reach, and using constraints to develop new capabilities.
10X your SEO with Semrush for Enterprise.
The world’s most powerful SEO platform, purpose-built for Enterprise.
Request demo
3 ways a stack responds to change and a career
Nassim Nicholas Taleb sorts systems into three groups by how they respond to stress. A fragile system breaks under stress. A robust system absorbs the hit and returns to where it was. An anti-fragile system comes out of the stress in better shape than it went in, like a muscle responding to training.
Fragile systems break under stress
A fragile stack depends on specific components staying exactly as they are. A vendor change, an API restriction, or a process the platform wasn’t designed to support can take the whole thing down.
A fragile career depends on one platform, one channel, or one manager staying in place.
Robust systems absorb the hit
A robust stack has backups and documented replacement processes, so it keeps running while repairs happen.
A career built to absorb shocks has a network, so it can recover from a layoff or a reorg to roughly its previous level.
Anti-fragile systems gain from stress
An anti-fragile stack goes further. It’s designed from the start for agility, replaceability, fit with the marketing processes it serves, and learning. Every change forces a review.
- Is the capability still needed?
- Does the process it supports still make sense?
- Is there a better option?
- Could the architecture be simpler?
The disruption produces learning, and the learning produces a better system than the one that was stressed. The same review applies to the processes, data, and people around the stack, which is where the career question begins.
The mechanism behind that third category is optionality. In practice, that means keeping more ways to reach an outcome than any single change can remove.
An anti-fragile career picks up a capability from each disruption that it didn’t have before because it’s built on properties that no single change can take away. Three of those properties matter most for a stack, and each one has a direct career equivalent.
Route optionality or the replaceability test
The first property is architectural. When you evaluate marketing technology, replaceability deserves as much weight as feature completeness, if not more.
Strong APIs, standard data formats, and portable data let you swap a component without rebuilding the system around it. Feature-rich platforms with proprietary lock-in look efficient until you need to leave.
Gartner has predicted that organizations that adopt a composable technology approach will outpace competitors by 80% in the speed of implementing new features, putting a number on what replaceability is worth.
Marketing expertise splits along the same line. Knowing how to configure one vendor’s journey builder, one platform’s audience tool, or one suite’s reporting layer is expertise tied to that vendor’s product decisions. When companies and enterprises migrate, that knowledge may not travel as well. Knowing how customer data should be modeled, how decisioning logic should be written to survive a platform change, how consent should flow through the stack, and how integrations should be designed is expertise that travels to any vendor because it describes the system rather than the tool.
AI raises the cost of getting this wrong. Agents are getting good at the operating layer — the configuring and button-pushing that used to be the visible part of a martech role. What they need from a marketer is the logic behind those actions. Someone has to decide what a good outcome looks like, what data the decision may use, and which rules can never be broken. That work sits above any single platform. It’s also the part of the job an agent can’t take over because it’s the part the agent runs on.
The test for your own capabilities is the same one I apply to a platform: Could you move the capability to a different vendor within a quarter? If yes, it’s portable, and it appreciates as platforms churn. If no, it could belong to the vendor, and you’re renting it.
Owned reach: The career version of first-party data
The second property is ownership. Zero- and first-party data are the clearest examples in a stack. When customer data lives inside an advertising platform, you have no options when that platform changes its terms. When it lives in a system you control, you choose how to activate it across channels. A platform change becomes an inconvenience instead of a crisis.
The career equivalent is the reach you control.
- Rented reach comes from a role, a title, a platform certification, or a manager who speaks for you in rooms you’re not in.
- Owned reach comes from relationships you hold directly with the people in IT, legal, finance, and data who have to cooperate for any marketing system to work. It also comes from the shared vocabulary you’ve built with them and the documentation you’ve authored that other people rely on.
When the role changes, rented reach resets to zero. Owned reach comes with you.
Documentation deserves a specific mention because it’s an underrated form of owned reach in martech. A team that maps its technology architecture, what connects to what, and where data flows, makes better decisions about additions and replacements. Most teams haven’t done it.
The marketer who creates that map becomes the reference point for every decision that touches it. That position outlasts vendors and reorganizations because the map is needed regardless of who reports to whom.
Gain from stress or positive friction applied to your own job
The third property is what I call positive friction. The instinctive response to a constraint — a privacy regulation, a platform limitation, or a budget cut — is to treat it as an obstacle to work around. In practice, many constraints improve outcomes by forcing prioritization. They reveal capability gaps rather than creating them. Privacy rules didn’t create the difference between organizations that understood their customers and organizations that only tracked them. The rules exposed it.
The same exposure happens to people. Every constraint that hits a marketing organization runs a test on the marketers in it. An AI mandate from the executive team is the current version of that test.
Some marketers wait for the vendor to tell them what the mandate means. Others take the constraint apart in five steps:
- Identify what specifically can no longer be done.
- Check which practices the constraint eliminates that should have gone away anyway.
- Work out what it forces the team to prioritize.
- Look for the advantage available to whoever adapts fastest.
- Ask whether it points toward something more sustainable, regardless of whether the constraint lasts.
The marketer who runs that review on the AI mandate becomes the person leadership asks when the next constraint arises.
That’s what gaining from stress looks like at the level of your career. The constraint is the same for everyone in the organization, and the marketers who engage with it early tend to emerge from it with something new.
What to do this quarter
Every step below is a practice I recommend for marketing systems, turned inward to help marketing careers become anti-fragile.
Run the replaceability test on yourself. List the capabilities that currently justify your role, then mark each one as portable or vendor-bound using the one-quarter test above. The portable column is the one to grow.
Marketing teams that stay ahead of change give their people 5% to 10% of their time for developments that may become important before urgency makes them urgent. Apply that to yourself. Spend it on capabilities that describe the system rather than a tool. Data modeling, consent and privacy architecture, and evaluating whether an AI output is fit to act on are three that travel to any vendor.
Write the map if nobody has. If your organization doesn’t have a current document showing which systems hold customer data, how they connect, and where consent and identity flow, produce it. The work takes weeks, requires no budget approval, and puts you in every conversation about adding or replacing a component.
Place one small bet. Taleb’s barbell strategy, applied to marketing investment, combines a large position in things that hold value in almost any future with small positions in things that could pay off enormously in a specific one.
For a career, the large position is the portable capabilities above. The small bet is an emerging capability you don’t need yet. Right now, one obvious candidate is designing or running a useful workflow with an AI agent end-to-end, with your own evaluation criteria, so you have operating experience before your organization demands it.
Then apply what Taleb calls via negativa, the discipline of removing what would fail in multiple futures. For a career, that means stopping your dependence on the following:
- A single vendor certification as the center of your professional identity.
- One channel’s results as the record of your impact.
- Platform-provided attribution as your only evidence that your work moved the business.
- The annual planning cycle as the only time you reassess what you should be learning.
Finally, volunteer for the next constraint. When the next privacy change, budget review, or AI directive lands, take ownership of the review instead of waiting for it to be assigned. The review is where capability is built, and the person who runs it is the one who gains from it.
Change is the stable condition
The marketing organizations that define the next decade are the ones that built systems designed to improve through exposure to change, and careers can follow the same rule. The marketers positioned to gain from the AI transition are those whose expertise and working relationships travel with them from one platform to the next, and who run the review each time a constraint arises.
None of that depends on predicting which platform, regulation, or model wins. Marketers who work with martech already practice each of these skills on the stack. Turning them toward a career is the new part, and it’s the part that turns the next platform change into a capability you didn’t have before.
