Navigating AI in Marketing Operations

It's pretty wild that we're still hearing rumors about how AI is going to totally replace humans — and it's even coming from the AI companies themselves.

In June 2025, Anthropic CEO Dario Amodei went on record saying that AI was going to wipe out half of all entry-level white-collar jobs, and that unemployment could spike to 20% in the next 1-5 years.

These numbers seem grim, but consider the source. Anthropic's entire goal is to get people really excited about AI and push them to pay attention and invest in it. Their business literally depends on it. And, unfortunately, fear is often the strongest motivator, so this statement makes sense.

That said, we do think it's high time to get familiar with AI. It's absolutely going to augment our work — just like automation did a decade ago. Things used to be entirely manual, and then we automated work and freed up time. We'll see the same process with AI.

When we polled MarOps folks, we discovered that almost half of people (43%) are taking the first steps with AI, while 30% are gaining momentum and doing multiple implementations. Only 2% consider themselves AI leaders at this point.

So what's holding more people back from being more comfortable with AI?

We know people struggle with a few major AI implementation challenges: integration complexity, quality control, and legal compliance concerns. Our audience was split into roughly even thirds, with each segment citing one of these as their primary roadblock.

And these are all valid! It can be confusing to bring a new tool on and figure out how it works with your current tech stack, let alone verify that the quality AI produces is up to speed and that you've operating within legal boundaries.

However, you stand to gain so much if you do AI the right way. For example, it's common for us to hear from teams who used to spend six hours across three roles on one blog post. Once they get comfortable using AI, that cuts down to 1.5 hours for just one role.

That's a cost of roughly $0.20 in prompts, versus the former $300 in hourly rates.

Again, notice that humans are not completely replaced in the above example. You still have to be involved in the process, and AI definitely needs to be trained on what to do. No AI tool can fully understand all of the context and complexities that you do, straight out of the box.

Sara McNamara, Revenue Operations and GTM Strategy Lead at Vector, has over a decade of marketing ops experience. One of her top tips regarding AI implementation is to always verify everything.

AI hallucinates and makes up things that aren't real so, yes, use it — but always double check before sending anything to press. If you're just brainstorming, that's one thing, but you definitely don't want to trust AI for analytics or nuanced reasoning.

You should also try not to be intimidated by what others are doing with AI. It's easy to look around and then feel like you'll never catch up to where other people are at.

However, if you go into it thinking you should immediately be able to do everything you're curious about, you're only going to get discouraged. Remember that some folks have a ton of budget behind them funding their AI exploration, too; it's okay if that's not where you are.

Start small and battle-test AI. Do what you can, and vet your tools and outputs with trusted stakeholders or internal champions. Once you find things that work (or don't), keep experimenting from there.

Here are a few use cases to experiment with:

You can use AI for automated GTM research, including web scraping. This can be useful if you've got a big list of event attendees, and you want to whittle it down to find the best personas to target. Manually searching through the list would obviously be cumbersome, so enlist the help of AI!

Here's how to do it:

1. Use a tool like ChatGPT or Clay's Claygent to import a sheet of anonymized attendees.
2. Enrich attendees with any missing core data, like company domain.
3. Create a detailed prompt with plenty of context; see the video above for more details on how AI tools can help with the prompt creation as well.
4. Keep enriching with relevant data until you feel you can make an educated decision on who to invite!

AI agents are all the rage right now; they can take over semi- or fully-autonomous activities like support question responses or research documentation. No more having to answer the same questions over and over again!

Here's how to do it:

1. Log into ChatGPT or Zapier and create an agent.
2. Provide it with documentation links or files so it has context.
3. Create prompts to instruct the agent on how it should generally approach answering questions — for example, tell it what to say when it can't find an answer. You should instruct it on its objective, the intended communication style, and give it any other applicable rules.
4. Share the chatbot link with internal champions to test it out and continue adding to it over time.

Once you're done, you can share this chat as the first stop for people with general questions. This can save your team time and get customers answers quickly. You can also collect info on what questions the agent was unable to answer so you can provide feedback and more context to the agent so it's prepared next time.

Marketing teams spend a ton of time pulling content together, including thought leadership content. We know that C-suite execs are super busy, but they also require LinkedIn and other social content — so use AI to make that happen without a ton of extra work.

Here's how to do it:

1. Record a call in Goldcast Recording Studio or Content Lab. (You could also use this workflow for a webinar or event!)
2. Once the call or event concludes, use the platform to swiftly repurpose the recording into video snippets, social media posts, blog drafts, key takeaways, and more.
3. Optional: Set up a voice profile in the platform to apply different brand voices for different speakers.
4. Share posts and other assets with stakeholders for final edits and tweaks before publishing.

If these use cases are feeling intimidating, don't worry — you can start where you're at just by looking at what eats up your marketing team's time and evaluating whether automation or AI can help. You probably already own tools like Zapier that can get you started!

PS: Don't be scared to automate just segments of a process. This can still save a ton of time, and you can always work toward automating the entire workflow.

Companies — especially public ones — have to be very careful about public perception and protect customer data as much as possible. No one wants their personal identifiable information (PII) compromised, and your customers want to trust that you're only sharing data that's truly necessary.

AI is still pretty new, so we'll see regulations evolve, the same way we did with GDPR and CCPA. For now, work closely with your legal and compliance teams — or consultants, or whoever you have on deck to manage security. Talk about the risk tolerance involved with AI usage. You want to be as efficient as possible and keep up with competitors; how can you do that while balancing risk?

Here are a few more actions to take:

• Explore tools that already have AI embedded within them, like Clay or Zapier. When you use those systems, it exports some of the risk onto that vendor instead of you using ChatGPT and assuming the responsibility.
• Start a regular call or meeting across teams to discuss how AI is being used and start formulating standards and policies. You could even set up an AI council within your company for this purpose.
• Keep ethical considerations in mind. AI can definitely be biased because it's feeding off all historical information, which can stereotype and contains biases. If you use AI to create external content, be super prescriptive with your prompts so there's no way it can go off on a tangent.

We'll close with a straightforward, seven-step playbook to scale AI across your organization:

1. Create a centralized AI playbook.

Keep careful notes as you use AI about the ways you're using it, including the prompts you've tried, results, and the systems you're leveraging. It's much easier to do this as you go instead of trying to go back and remember things after the fact.

1. Set up a prompt library.

If you're feeling really fancy, go ahead and set up a prompt library so people don't have to reinvent the wheel. They can just go to your library and browse the available prompts to see if there's one that they can also apply to their use case.

1. Assign AI champions.

If you can find a lead within marketing to champion AI efforts and keep up with the latest developments, that will help drive adoption throughout the company. When people are excited about the results they're seeing through using AI, their colleagues are naturally going to want to get up to speed so they can do the same.

We just mentioned this, but AI councils can be useful because they give people a set group to ask if they have questions and task certain folks with keeping track of the AI playbook and relevant policies.

1. Standardize tools and access.

It's easy to duplicate efforts and even tools if you aren't communicating with other teams about what you're doing. Standardization requires open lines of dialogue and transparency, which is why those regular meetings across departments or establishing an AI council is so important.

1. Train the team.

The CEO of Zapier recently shared that they assess each employee on their AI fluency, which keeps everyone on the same page. If you're aiming for everyone to have the same level of familiarity with AI, people need to know where they stand and who they can go to for advice and tips —plus, by assigning marks, you can see how much training support your team needs.

1. Govern usage

Don't let folks run wild with AI; you don't want to end up with security issues or data leaks and negatively impact public perception of your company. Especially if you're first starting out, be clear about how people should and shouldn't be using AI.

1. Create feedback loops

Be sure to give folks a way to provide feedback. Set up a channel on Slack and invite comments and observations so you're quick to learn about hallucinations, missteps, and knowledge gaps. This is also how you can discover things AI is doing well at and improving.

• AI will not entirely replace humans. Instead, look to AI as a powerful augmentation tool that can save you time and resources and free you up for more complex tasks.
• Tip: Try one of the three provided use cases to see how AI can benefit MarOps research, customer support, and content production.
• Legal considerations abound in the AI space, and we'll likely continue to see regulations pop up as people become more familiar with AI and associated risks and concerns.
• Tip: Work closely with legal and compliance now to define risk tolerance and set up org-wide policies. Review the seven-step framework to scale AI and use it as a guiding light.