Why Guardrails and Markdown Files Could Define the Next Era of Marketing Operations
- AJ Navarro, Marketing Operations Manager at Sprout Social
- Account Based Marketing (ABM), AI and Automation, Campaign Optimization, Content Strategy and Marketing, Customer Experience (CX), Data Management and Analytics, Digital Transformation, Lead Gen / Lead Mgmt, Managing Org Transformation, Managing Your Career, Martech Stacks, Multi-Channel Journeys, Other Martech, Platform and Data Integration, Sales and Marketing Orchestration, Strategy and Trends
As AI agents become more common in marketing operations, the biggest opportunity may not be automation alone. It may be governance.
In a recent conversation with Martechify, AJ Navarro, Marketing Operations Manager at Sprout Social, argued that the next major shift in martech will not simply come from faster AI tools or smarter workflows. Instead, the real differentiator will be how well teams build guardrails, document rules, and train AI systems using structured instructions such as Markdown files.
For marketing operations teams, that is a critical shift. The conversation around AI often focuses on productivity gains, time savings, and new capabilities. AJ offered a more grounded perspective. He said that as AI agents become more embedded in day-to-day operations, companies will need a much stronger foundation to keep those systems accurate, controlled, and useful.
“Every organization, as the industry shifts, is going to suffer from two problems, which are governance and guardrails, AJ said. That observation framed one of the most important discussions in marketing operations right now. AI can accelerate work, but without clear instructions and constraints, it can also create confusion, inconsistencies, and costly mistakes.
In a recent conversation with Martechify, AJ Navarro, Marketing Operations Manager at Sprout Social, argued that the next major shift in martech will not simply come from faster AI tools or smarter workflows. Instead, the real differentiator will be how well teams build guardrails, document rules, and train AI systems using structured instructions such as Markdown files.
For marketing operations teams, that is a critical shift. The conversation around AI often focuses on productivity gains, time savings, and new capabilities. AJ offered a more grounded perspective. He said that as AI agents become more embedded in day-to-day operations, companies will need a much stronger foundation to keep those systems accurate, controlled, and useful.
“Every organization, as the industry shifts, is going to suffer from two problems, which are governance and guardrails, AJ said. That observation framed one of the most important discussions in marketing operations right now. AI can accelerate work, but without clear instructions and constraints, it can also create confusion, inconsistencies, and costly mistakes.
Watch the full interview with Sprout Social’s AJ Navarro to hear why governance, not just automation, will define the next era of marketing operations. Learn how guardrails, Markdown files, and structured documentation are reshaping how teams train AI systems and deliver reliable results.
AI is moving fast, but governance is not keeping up
AJ said many teams are already excited about AI agents that can help with list management, lead investigation, reporting, and workflow support. He described several promising use cases, especially within Adobe’s newer AI features, but he repeatedly returned to the same concern: Teams are not talking enough about the operating structure behind those agents.
According to AJ, companies are paying attention to what agents can do, but not always to how those agents should be governed.
He noted that this gap matters even more in marketing operations because the function often sits at the center of data quality, process integrity, and system logic. If AI tools begin making or influencing operational decisions, then the quality of their instructions becomes just as important as the quality of the data they access.
That is where Markdown files come in.
According to AJ, companies are paying attention to what agents can do, but not always to how those agents should be governed.
He noted that this gap matters even more in marketing operations because the function often sits at the center of data quality, process integrity, and system logic. If AI tools begin making or influencing operational decisions, then the quality of their instructions becomes just as important as the quality of the data they access.
That is where Markdown files come in.
Why Markdown files matter
AJ said one of the most overlooked skills for marketing operations practitioners is learning how to understand, write, and maintain Markdown files, often referred to as MD files. “We as marketing operations practitioners need to get very good, if not almost perfect, at understanding, writing, and reading Markdown files, or MD files,” AJ said.
He explained that in some AI-driven environments, Markdown files can act as instruction layers that tell an agent how to behave. That can include naming conventions, workflow rules, acceptable outputs, process documentation, and what the system should do when information is missing or unclear. Instead of repeating prompts manually every time, teams can define those standards in a structured file and use them as a governance framework.
As AJ described the new reality of marketing operations, “Something my team is going to start working on is becoming really good at MD file structures. Those can be anywhere from one page to 20 pages. They can be very complex or very simple. The best way I could put it is this: If you are an organization with all your weekly documentation—this is our lifecycle, this is our lead score—take all of that, because that is your new area for training these agents and systems. It is going to be very important for governance and guardrails.”
That simple idea has major implications. In the past, marketing operations teams focused on campaign logic, scoring models, routing rules, and reporting frameworks. Going forward, they may also need to manage the instruction sets that shape how AI tools interpret and execute those systems.
He explained that in some AI-driven environments, Markdown files can act as instruction layers that tell an agent how to behave. That can include naming conventions, workflow rules, acceptable outputs, process documentation, and what the system should do when information is missing or unclear. Instead of repeating prompts manually every time, teams can define those standards in a structured file and use them as a governance framework.
As AJ described the new reality of marketing operations, “Something my team is going to start working on is becoming really good at MD file structures. Those can be anywhere from one page to 20 pages. They can be very complex or very simple. The best way I could put it is this: If you are an organization with all your weekly documentation—this is our lifecycle, this is our lead score—take all of that, because that is your new area for training these agents and systems. It is going to be very important for governance and guardrails.”
That simple idea has major implications. In the past, marketing operations teams focused on campaign logic, scoring models, routing rules, and reporting frameworks. Going forward, they may also need to manage the instruction sets that shape how AI tools interpret and execute those systems.
Why Adobe’s AI agents could change admin work
AJ expanded on Adobe’s newer AI capabilities, particularly tools still being tested in beta and alpha environments. While he said not all of them are production-ready yet, he sounded especially enthusiastic about two use cases: list management and lead investigation.
On the list management side, AJ said, “That list management one—that one’s just a game changer.” For organizations handling large event volumes and messy real-world data, he noted that AI could reduce manual cleanup and ease pressure on already stretched teams.
He was equally interested in the lead investigation agent, which could help administrators understand why a lead did or did not qualify as a marketing lead (MQL). Instead of manually checking multiple programs, records, and activity histories, that can take an admin between “one to four hours” according to AJ’s estimates, the agent could pull together an explanation in seconds.
That efficiency matters, but AJ did not present it as a reason to lower standards. If anything, he used these examples to reinforce his larger point. As AI agents take on more operational work, governance becomes even more important. Faster outputs are only valuable if the underlying rules are correct.
In other words, Adobe’s AI agents may change admin work by speeding it up, but Markdown files and guardrails will determine whether that speed produces reliable results. “We don’t always want to blindly trust AI,” he said, “but it gives us time back into our day.”
On the list management side, AJ said, “That list management one—that one’s just a game changer.” For organizations handling large event volumes and messy real-world data, he noted that AI could reduce manual cleanup and ease pressure on already stretched teams.
He was equally interested in the lead investigation agent, which could help administrators understand why a lead did or did not qualify as a marketing lead (MQL). Instead of manually checking multiple programs, records, and activity histories, that can take an admin between “one to four hours” according to AJ’s estimates, the agent could pull together an explanation in seconds.
That efficiency matters, but AJ did not present it as a reason to lower standards. If anything, he used these examples to reinforce his larger point. As AI agents take on more operational work, governance becomes even more important. Faster outputs are only valuable if the underlying rules are correct.
In other words, Adobe’s AI agents may change admin work by speeding it up, but Markdown files and guardrails will determine whether that speed produces reliable results. “We don’t always want to blindly trust AI,” he said, “but it gives us time back into our day.”
The human in the loop now includes the file in the loop
One of AJ’s strongest points was that troubleshooting will change. In a traditional marketing automation environment, when something breaks, an admin might inspect a person record, activity history, smart campaign, or sync logic. In an AI-supported environment, part of that investigation may shift to the Markdown file itself.
AJ explained that if an AI agent produces a flawed answer or takes the wrong action, teams will need to ask whether the issue came from the data, the workflow, or the instructions that guided the agent. That means governance documents are no longer static support materials. They become part of the system.
“One of the first places I need to be checking is going into that MD file,” AJ said. That is a meaningful evolution for marketing operations. Documentation has often been treated as a nice-to-have. AJ’s view suggested that documentation, especially in structured formats used by AI systems, is becoming a core operational asset.
He also stressed that teams need to be explicit in how they write these instructions. If an agent does not know the answer, the file should make clear that the agent should not invent one. “If you do not have the answer, do not provide the answer,” AJ said.
That kind of rule may sound simple, but it addresses one of the most persistent problems with AI: hallucination. Without clear boundaries, an agent may produce something that sounds plausible but is completely wrong. In a marketing operations context, this can affect reporting, lifecycle stages, lead handling, or customer-facing experiences.
AJ explained that if an AI agent produces a flawed answer or takes the wrong action, teams will need to ask whether the issue came from the data, the workflow, or the instructions that guided the agent. That means governance documents are no longer static support materials. They become part of the system.
“One of the first places I need to be checking is going into that MD file,” AJ said. That is a meaningful evolution for marketing operations. Documentation has often been treated as a nice-to-have. AJ’s view suggested that documentation, especially in structured formats used by AI systems, is becoming a core operational asset.
He also stressed that teams need to be explicit in how they write these instructions. If an agent does not know the answer, the file should make clear that the agent should not invent one. “If you do not have the answer, do not provide the answer,” AJ said.
That kind of rule may sound simple, but it addresses one of the most persistent problems with AI: hallucination. Without clear boundaries, an agent may produce something that sounds plausible but is completely wrong. In a marketing operations context, this can affect reporting, lifecycle stages, lead handling, or customer-facing experiences.
From documentation to strategic skill set
One of the most interesting implications of AJ’s perspective is that Markdown literacy may soon become a real career advantage in marketing operations.
For years, the most valued skills in the field included platform expertise, campaign building, lead lifecycle management, integrations, analytics, and stakeholder communication. Those skills still matter. But AJ suggested that another layer is emerging: the ability to translate business rules into structured AI instructions.
That is a different kind of operational fluency. It requires teams to think not just about what they want a system to do, but how to define acceptable behavior in advance.
AJ said Markdown files can be as simple or as complex as needed. A team might use them to document naming conventions and program structure, or to encode broader lifecycle logic, governance policies, and decision-making rules. Either way, he said, they will play a growing role in how organizations train and constrain AI agents.
For marketing operations professionals, that means documentation is becoming more strategic. The people who can build clean processes and express them clearly may be the ones best positioned to lead in an AI-first environment.
For years, the most valued skills in the field included platform expertise, campaign building, lead lifecycle management, integrations, analytics, and stakeholder communication. Those skills still matter. But AJ suggested that another layer is emerging: the ability to translate business rules into structured AI instructions.
That is a different kind of operational fluency. It requires teams to think not just about what they want a system to do, but how to define acceptable behavior in advance.
AJ said Markdown files can be as simple or as complex as needed. A team might use them to document naming conventions and program structure, or to encode broader lifecycle logic, governance policies, and decision-making rules. Either way, he said, they will play a growing role in how organizations train and constrain AI agents.
For marketing operations professionals, that means documentation is becoming more strategic. The people who can build clean processes and express them clearly may be the ones best positioned to lead in an AI-first environment.
What this means for marketing operations leaders
AJ’s message was clear: The future of marketing operations will not be defined by AI adoption alone. It will be defined by whether teams can implement AI responsibly.
That responsibility starts with guardrails. It continues with structured documentation. And it depends on people who understand both the technical systems and the business consequences when those systems go wrong.
In that sense, Markdown files are more than a tactical detail. They represent a broader shift in how work gets operationalized. They are becoming part policy, part prompt, and part process map. For teams adopting AI agents, they may become one of the most important tools in the stack.
That responsibility starts with guardrails. It continues with structured documentation. And it depends on people who understand both the technical systems and the business consequences when those systems go wrong.
In that sense, Markdown files are more than a tactical detail. They represent a broader shift in how work gets operationalized. They are becoming part policy, part prompt, and part process map. For teams adopting AI agents, they may become one of the most important tools in the stack.
AJ’s advice for the next generation of martech professionals
Toward the end of the conversation, AJ offered advice for people entering marketing operations and martech today. His guidance was practical and refreshingly grounded in both community and mindset.
First, he said, “Dive right in and find a community.” He credited community support with helping shape his own career and suggested that newcomers should surround themselves with peers, practitioners, and mentors who could help them learn faster.
Second, he encouraged people to stay flexible as tools and practices change. In a field where workflows can shift in a matter of months, the ability to keep learning matters as much as any single certification or platform skill. As he put it, “Always have the mindset of a student.”
He recalled that one of the best CEOs he has ever worked for told him during a mentorship talk, “We don’t hire for people’s skill sets. I didn’t hire you just because you knew Marketo. We focus on people’s ability to solve problems.”
Finally, AJ emphasized the importance of soft skills. Technical knowledge matters, but the ability to communicate clearly, collaborate with stakeholders, and connect operational work to business outcomes is often what separates strong practitioners from future leaders. In an industry evolving as quickly as martech, curiosity, adaptability, and problem-solving may ultimately become the most valuable skills of all.
First, he said, “Dive right in and find a community.” He credited community support with helping shape his own career and suggested that newcomers should surround themselves with peers, practitioners, and mentors who could help them learn faster.
Second, he encouraged people to stay flexible as tools and practices change. In a field where workflows can shift in a matter of months, the ability to keep learning matters as much as any single certification or platform skill. As he put it, “Always have the mindset of a student.”
He recalled that one of the best CEOs he has ever worked for told him during a mentorship talk, “We don’t hire for people’s skill sets. I didn’t hire you just because you knew Marketo. We focus on people’s ability to solve problems.”
Finally, AJ emphasized the importance of soft skills. Technical knowledge matters, but the ability to communicate clearly, collaborate with stakeholders, and connect operational work to business outcomes is often what separates strong practitioners from future leaders. In an industry evolving as quickly as martech, curiosity, adaptability, and problem-solving may ultimately become the most valuable skills of all.
Related resources
Want more like this resource? Find helpful links below.
