OpenAI’s marketing head shares the one big mistake CMOs are making with AI

When Krithika Muthukumar, head of marketing for OpenAI, stepped into her role last August, she hardly needed to worry about the top of the funnel—aka building awareness—for the brand.

The hype around the company’s generative AI products like ChatGPT and DALL-E had already reached stratospheric heights at the research organization. By then, more than 1.4 billion people were visiting its site, and its OpenAI Enterprise business was set to launch with a waitlist of companies outside its virtual gates. Despite very little traditional “marketing,” the world’s leading AI company reportedly reached a valuation of $80 billion three months ago.

Still, Muthukumar, who started her career as a front-end developer at Google before moving into marketing at companies like Stripe and Dropbox, faces other heady challenges. For one, she has to nudge consumers and OpenAI’s enterprise customers—the company declined to share how many they have signed up—to get past the novelty factor of AI. (More on that below.) And let’s not forget that OpenAI Enterprise has to protect its turf from competitors, many of whom offer cheaper enterprise tools.

Then there’s the moral responsibility Muthukumar says she and others at the organization have taken on as part of their daily work: OpenAI’s stated mission is to help humanity understand AI and solve its many ethical issues. That noble goal was a big part of the draw when Open AI recruiters contacted her, she explains, calling the invitation “unignorable.”

The job also ticked another box for her. “The intersection of highly technical products and marketing was really my passion,” she says. “I took that passion to be the first marketer at Stripe and saw that company scale from around 60 people all the way up to 7,500 people and scale the marketing team accordingly to hundreds of people around the globe.” She’s repeating that act at OpenAI, rapidly adding to her team in a competitive environment for those with AI expertise.

In her role, Muthukumar says she not only pitches OpenAI products but also educates companies and fellow CMOs about what she sees as AI’s transformational power when it’s applied strategically with a view to long-term impact.

This interview has been edited for length and clarity.

OpenAI already has a large following around the world, so how are you marketing its products? 

In the consumer world, I would say that everyone is now aware of AI, but folks aren't aware of how best to use it and how it's going to fit into the fabric of their lives. If we want to engender use-case epiphanies for folks, that is a different challenge. And we’ve approached that by sharing a lot more examples. You'll notice on our Instagram channels recently, we've been sharing small videos and tidbits. We're carrying that through to the business side, too.

We’re also thinking about the responsibility that we have in marketing not just our products and technologies but AI as a whole. That's something we are tackling across a multitude of functions. Our policy teams reach out to governments and educational institutions. Our communications team is working on more broad-based AI education, and in the marketing team, we have to think about, for example, the way you interact with a chatbot to get the best results. 

What are business leaders asking you about your OpenAI products? Are there common misconceptions you need to clear up?

With many of the folks we talk to, it feels like they have this mandate and urgency to bring AI into their organizations. Often, our teams ask them, “What are you trying to do with it?” It's not enough to check the box and deploy an AI tool into your organization. It's important for you to critically identify the tasks and workflows that are taking up considerable time for employees and how to make those tasks a lot easier. We see that approach as a strong marker between folks who have very successful AI deployments and folks who don't.

There also still isn’t enough storytelling about those durable use cases that you can build with AI that end up being revenue drivers for your company or things that increase your customer engagement. For example, we've seen Moderna adopt AI to help with everything from legal reviews to reducing the time it takes to calculate the right doses for some of its pharmaceuticals.

How do you see AI changing the CMO job in the next year or two?

In the marketing discipline, there are a lot of repetitive things, like time-consuming workflows, that go beyond content and aren't being embraced by AI right now. Think about all of the manual work that is being done today by downloading spreadsheets from multiple different systems and having to stitch them together to get a full view of your funnel. AI can combine that data and unlock much more personalized marketing that is very specific to the person who's coming to your site.

AI can also be a fantastic way to round out marketers with specific strengths and expertise. For example, if you have a very creative person who is thinking about brand campaigns, how can you augment their skills with more of the data-driven and data science side that AI can provide? Similarly, if you have a very technical marketer, are there things that an AI chatbot can help them do to think more creatively and be a brainstorm partner?

Looking further ahead, will AI eventually take away CMO jobs? We've already seen headlines about AI doing the work of hundreds of employees.

No, at least not from the conversations I've been having with CMOs.

Our ethos has always been about iterative deployment at OpenAI. We don't think AI and surprise go well together. People have seen previews of the capabilities to come on an ongoing basis, and we expect to continue that same model.

AI is also much more geared toward replacing tasks. Today, it can handle five-second or five-minute tasks. We expect it'll be able to take on five-hour or maybe even five-day tasks on the more distant horizon. That frees up your workforce to be much more creative, focus much more on the strategic challenges, and unlock entirely new ways of marketing.

How are CMOs experimenting with generative AI? What advice do you give them?

Many of their team members are already using ChatGPT for content creation or to help write professionally, whether that’s strategy memos or docs for internal consumption. But it has to be more of a top-down approach, with CMOs saying, “Hey, this is how we're going to embrace AI as an organization,” versus waiting for folks within the organization to raise their hand and see what use cases come up.

CMOs should consider integrating AI across a multitude of marketing functions, starting with understanding their data and customer insights. They should also consider creative, more meaningful, and personalized ways to think about going to market, and finally, measure the impact of marketing and hold AI accountable. These are all places where AI can have an impact, but it requires systems and data investment to make it more impactful than just being a writing assistant.

People also need to think on a much longer time horizon in terms of embracing AI within their marketing teams. Today, it feels like it’s treated as a one-quarter problem.

You must have found a creative or unexpected use for AI at work.

ChatGPT is the best data analyst I've ever had. I can slice and dice and do funnel analysis for as long as I'd like and ask dumb questions to get a more grounded understanding of the data. In fact, when we launched ChatGPT Enterprise, our lead volume shot through the roof. I was able to pair those leads with ChatGPT to create our first lead scoring model, which we actually applied to prioritize and route the most important prospects to the sales team.

Have you noticed that some marketers are already pulling ahead in their AI adoption—and what are they doing?

One of the biggest changes they've made is integrating custom ChatGPT data. Whether it's Salesforce data, an internal document source like Notion, Google Docs, or Microsoft Office 365, they’re using AI to combine data sources. If you have a strong understanding of your customer, then the AI can provide much more interesting information at the other end.

The same goes for data analysis. I've used ChatGPT myself to do funnel analyses before we had a marketing operations person in-house.

What is your favorite use of ChatGPT or any of the OpenAI tools? 

On the business side, one of my favorite use cases is from a customer working with about 400,000 farmers in India and Kenya. This company, Digital Green, is working to empower farmers with information and data to help them increase the yield of their crops. They’re using a chatbot as an assistive approach to reach these farmers in their local languages. It has increased farmer income considerably.

On the personal side, it has been fantastic for me to generate custom coloring pages for my nieces and nephews. It's the best party trick. You tell Dall-E to generate coloring pages based on their interests, and it takes up hours of their time, and they love it.

This story was originally featured on Fortune.com