There are a few universal questions that keep CMOs up during the quiet hours of the night: How do we do more with fewer resources? And how do we make a greater impact without burning out our teams?
When Lisa Cole, Chief Marketing Officer at Cellebrite, stepped into her role last year, these were the questions constantly on her mind. She wanted to illustrate the marketing department’s ability to drive results and improve work-life balance for employees.
Enter the impetus for solving both challenges: artificial intelligence (AI).
Cole brought in Nicole Leffer, a CMO AI advisor, to help navigate this new frontier. I spoke with them to learn how marketing leaders can incorporate AI across their teams effectively. Their partnership is just one example of how brands are using AI in marketing to achieve incredible results.
“Rather than fight against technology, how do you train and develop the next generation of marketers so they can leverage it to have a greater impact on the company—rather than be disrupted by it? It feels better to take control,” Cole says.
Tapping the marketing strategist in your pocket
Both Cole and Leffer have interesting philosophies when it comes to AI. For Cole, AI is an ever-present partner to brainstorm and ideate with.
“Oftentimes, I get my best ideas during nights and weekends, when I don’t want to bother my team. But I also don’t want to slow down. So when I want to flesh out an idea, the world’s smartest marketing strategist is in my pocket at all times,” she says.
Cellebrite is a mission-driven organization whose technology is used to save and protect lives, so they wanted to mobilize a movement. When Cole was brainstorming for a new publishing commitment campaign, she looked to ChatGPT to help pressure test and anchor for the campaign. She says this critical brainstorming happened on a rainy Saturday afternoon at home. She asked a variety of questions, from solutions to fighting crime to how to increase global safety.
“I was trying to research websites and data points, but I wasn’t getting there fast enough. [AI] helped me connect dots faster than I could have done on my own. There were some real themes that came out of that exchange. And I used it to flesh out what became the framework for a global campaign,” she says.
With Cole’s personal experimentation being so successful, she was motivated to bring Leffer on board to illustrate the power of the tool across her team.
Combating challenges and nurturing exploration
Leffer encourages clients to embrace an experimental mindset to overcome the many challenges of implementing AI, from combating learning curves to crafting better AI prompts.
A common pattern emerges when marketing teams begin to adopt the technology. Many people begin experimenting with AI because they’ve heard about the hype, or they’re skeptical and want to learn more. In some cases, good first impressions are wiped away when users start discovering issues like hallucinations and glitches.
“Generative AI hasn’t worked out all the kinks. It forgets things every now and then. So beginners start running into the errors or they don’t get the results they really want, especially if they don’t know to prompt correctly,” she says.
She explains there are groups of people who will step away once they hit an error, but there’s also a smaller segment of users who will have a more experimental mindset.
“Once you start experimenting, you start seeing how to overcome those limitations. The more someone experiments, the more they learn. You get to the point where you’re like me or many of my clients—you’re using AI constantly. I’m always trying to see what’s possible.”
Building a culture of experimentation
In her experience training CMOs, Leffers says the most common challenge is getting teams to embrace and use new technology. Ultimately, it comes down to leading by example. Executives must shape a culture of using and studying AI.
“You can’t just give the people the tool once, and then expect them to adopt it. Certain people are going to get so excited, and they’re going to go run with it. But other people need to be reminded. You’re changing habits that they’ve had their entire lives and professional careers,” she says.
Leffer notes there will be generational challenges as well. She says many people think younger employees will adopt AI fast while older generations will need more time, but it’s actually often the opposite.
“Anyone who has blown in a Nintendo cartridge to fix it, grew up having to tinker with technology to get it to work. The technology wasn’t already ready for you. You learned to navigate these technologies without a guide. I think those people are having an easier time adopting AI. Younger generations have had an iPhone most of their entire life. There was no figuring it out because it was ready to go. So adopting something as open-ended as AI is harder,” she says.
Data from The Sprout Social Index™ reinforces this phenomenon, as digital natives are most likely to be concerned about the emergence of AI in social media interactions. Some 46% of 18-24-year-olds say they’re apprehensive about increased AI usage, making them second only to consumers ages 57-75.
Proving the power of AI in real time
Cole’s advice for overcoming the challenges of incorporating AI across various teams and generations? Practice what you preach and show others the best way to tackle the technology.
“What has worked for my team is to prove to them—with real evidence—that the output reflects the quality of input that you put in. [Showing them] it’s meant to be an iterative process versus prompting AI and using the initial response as your final product,” she says.
Showing these proof points when collaborating with Leffer helped Cole’s team see the power of AI. Cole gave Leffer their most common workflows, personas, messaging framework, and their brand voice and tone guide. Leffer used these foundational inputs to create real examples of how the team could use AI. For example, Leffer produced a blog article and a series of emails to promote the piece and other distribution assets.
“She walked through how she got there in real time. We proved to them the output could be really strong. Then we provided them with the prompts and the training on the iterative process to question it, to strengthen the end result,” Cole says.
The art of crafting the right prompt
Leffer underscores how part of the AI learning curve stems from not knowing how to prompt. Instead of chatting back and forth and just asking/answering questions, she recommends starting an initial query and using ChatGPT’s prompt edit button to reflect the difference in the desired output.
“I’ve learned how to prompt so much faster because you see directly what information it needs, what’s irrelevant and what changes the outputs. Early on, it might’ve taken me six to eight edits to get what I wanted, whereas now, one or two will get me there,” she says, “I hear other people talking about how they had too long a chat, so ChatGPT started forgetting. You don’t have that issue when you’re refining through the edit button.”
Cole agrees this iterative approach is necessary for refining and differentiating a point of view or message. She explains that when she uses ChatGPT, she’ll refine outputs by asking for clarification, alternatives or to edit for brevity.
“It’s a conversation. It’s almost like a music composer. They might hear the same chords, but the way they put the chords together, the music itself, that’s a reflection of you and I bantering and brainstorming,” she says.
5 steps for incorporating AI across your marketing organization
Here are five steps for incorporating AI into your teams, based on Cole’s and Leffer’s advice:
1. Encourage failure
Leffer advises cultivating a culture where failure is OK. She would rather teams experiment and fail than not try at all. She recommends celebrating when people use AI and sharing those tests across the team.
“Recognize that people on your team come from different backgrounds and comfort levels. This is an opportunity to elevate everyone to an even playing field. But it’s also another place where we need to make sure people aren’t slipping through the cracks,” she says.
Leffer recommends going beyond the common approach of asking “Where can I use AI, or what things can I do with AI?” Instead, she recommends reversing this philosophy and ask, “Can I use AI for this? How?”
She advises using AI as often as you can to accelerate the learning process. Instead of knowing how to use AI for one or two things, you open the door for wider adoption.
2. Identify opportunities to use AI in current projects
Leffer advises leaders to ask about AI in team discussions to help teams understand how this new resource connects to their day-to-day work.
“One thing I found really helpful with getting my team to adopt the technology was, whenever we would be talking about projects, I would immediately ask, ‘How are you going to use our AI tools for this?’’’ Leffer says.
You can also share prompts and best practices on your internal communication channels to build a culture where everyone is expected to play around with AI.
3. Break down workflows step by step
When it comes to improving workflows, Leffer advises teams to first audit every discrete step in their existing processes. Identify where AI can expedite your workflow or improve the quality of your final output.
“Unless you’re sure the tool is going to extend something that takes 10 minutes to three hours, try to incorporate AI. You might find that you didn’t think AI would make a big difference. But if it saves you 15 minutes 20 times a day, you’ve just saved a lot of time,” she says.
She referred to the example of writing a blog, which encompasses a content brief, research, drafting and reviews. After reading the content brief, the writer starts researching. From there, the writer could feed content from their own research into ChatGPT, perhaps to organize key bullet points. Then you can continue leaning on AI to craft an outline or help you work on the first draft.
You have to be careful about doing research with any generative AI tool. Sometimes they will present information for illustrative purposes or will hallucinate and claim something is factual when it actually isn’t. Hallucinations can stem from the AI erroneously connecting inputs to another idea. Whether you’re using AI for social copy, video scripting or event collateral, fact-checking is vital. AI is not a research or creative replacement—humans should still review and build on anything coming out of these tools.
For example, when Cellebrite had to rename one of its products, a cross-functional team began brainstorming and thinking through how to defend their options. Each person used ChatGPT individually for ideas. Once the team was aligned on the best name, but before going into legal vetting, the group asked ChatGPT why the frontrunner was better than the others. Cole reminisces on how excited the team was to get a creative break and feel confident about the decision because they could articulate why the name was the right choice.
“AI reinforced the team’s collaboration and got them to a solution faster than several meetings would have. Increased speed to market and improved collaboration has been our biggest benefit of incorporating AI,” Cole says.
4. Be clear with what AI can be used for
Don’t let fear get in the way of people exploring AI. CMOs should partner with leaders across the business and consult with their legal counsel to develop an AI use policy.
“I see a lot of marketers who are hesitant to use it, because they don’t know what they’re allowed to do or not do,” Leffer says, “No one wants to feel like they’re sneaking around or doing something wrong. Make it clear on what is allowed, welcomed and encouraged.”
As Cole continues to work with AI, she’s concerned about what information is included in prompts, especially when it comes to protecting proprietary company data.
“I think about how we’re managing the data inputs and making sure we’re not putting anything sensitive on the other side. It’s important that we’re validating what we’re using, crediting the source and ensuring that the final output is compelling and differentiated,” she says.
Along with verifying outputs, leaders and teams should closely monitor the evolving ethics of AI.
“AI can enable us to do things we probably shouldn’t do, and we know that we shouldn’t do it. For instance, someone might scrape [a competitor] website or social media channels to use certain information against them. The red flags that your gut checks for should still apply,” she says.
5. Offer ongoing skills training and resources
Don’t assume people will figure it out on their own. Give them development resources that are tailored to your marketing teams’ specific roles and disciplines.
“If they’re a social media writer, give them resources around how to use AI for social media content. Talk through the use cases that are most relevant so they can see how to apply it,” Leffer says.
Take advantage of marketing communities like The Arboretum that connect professionals with their peers in real time so they can learn and explore together, especially when it comes to figuring out how to fit AI into their daily processes.
Preparing the next generation of marketers
We’re only in the early days of understanding the value AI can bring to marketing teams, with leaders like Cole and Leffer paving the way.
Today, Leffer says the biggest benefit of AI is the efficiency gain. “It opens up the potential to take on more projects, do things you maybe wouldn’t have had time for, and use your thinking for other higher level strategic work. That efficiency gain leads to being able to do more, which leads to a revenue gain at the end of the day,” she says.
To learn more about why over 80% of marketers say artificial intelligence (AI) has positively impacted their work (and how they plan to use it going forward), download The Sprout Social Index™.
How to use AI writing prompts to get the best out of your AI toolsPublished on October 31, 2023 Reading time 9 minutes
Top AI use cases in marketing to elevate your 2024 strategyPublished on October 19, 2023 Reading time 8 minutes
How to craft an effective AI use policy for marketingPublished on September 26, 2023 Reading time 9 minutes
A marketer’s guide to natural language processing (NLP)Published on September 11, 2023 Reading time 8 minutes