The Definitive CMO Guide to Artificial Intelligence


A CMO has a much harder job today than ever before. Digital has changed the environment, bringing both challenges and opportunities along with it. Marketers now have data to track consumer purchasing habits and the efficacy of marketing efforts. In 2016, “marketing budgets increased to 12% of company revenue…and CMO marketing technology spend is on track to exceed CIO tech spend in 2017.” As a result, executives are starting to take marketing more seriously and giving CMOs more responsibility. Along with a seat at the executive table and bigger budgets comes the expectation to track and prove return on investment. But predicting the success of campaigns is still difficult.

Marketers are up against an increasing number of media, the daily onslaught of new content, and the unpredictability of the consumer journey. To be successful, you must mine data to uncover insights, identify future trends, and intelligently apply learnings to more accurately predict the outcomes of future strategies. The methods for doing that are: work closely with a data scientist and adopt technology to supplement the humans.

Advances in big data, machine learning, and artificial intelligence (AI) have made those technologies the more practical, beneficial investments. The amount of data available makes it impossible for a human to analyze and make decisions in real-time. Instead, industry leaders are investing in artificial intelligence to make things that have pained marketers for decades possible. AI is a useful tool to uncover insights more quickly, turn those insights into guidance for future campaigns, and apply the learnings for better results, faster.

What is AI?

Most people still think of AI as robots, like in the movies “Ex Machina” and “Her”, or the “hosts” in the HBO series “Westworld.” Those androids have “artificial general intelligence,” and can perform more general activities, as the term implies. However, artificial intelligence can be applied in many other areas in more specific, though arguably more impactful, ways. Techopedia defines artificial intelligence as “an area of computer science that emphasizes the creation of intelligent machines that work and react like humans,” but those machines are not necessarily autonomous androids.

Most likely, you’ve already seen and benefited from AI in action. If you have ever typed a word into Google to find its proper spelling, you have used machine learning (invented in 1959 as a step towards artificial intelligence). Product recommendations, like Amazon or Netflix; virtual assistants, like Siri, Cortana, and Alexa; connected devices, like the Nest thermostat; and so much more, use artificial intelligence to make our lives easier.

Artificial intelligence can be applied in almost every industry to streamline processes. The reason AI is so diverse in application is because, as MIT labor economist David Autor explains, computers are “symbolic processors that follow codified sequences of instructions, programs or rules.” Any processes that “have explicit procedures for accomplishing them…routine tasks [that] can be codifiable” can be accomplished using artificial intelligence.

Why would we continue to perform tedious tasks that take up lots of our time, which we could better spend on the things computers will never be able to do?

Regarding marketing, AI automates, optimizes, and measures the repetitive marketing activities, saving time for more strategic, creative work. Any task you do repeatedly, whether it’s onboarding new users, defining customer segments, or posting to Facebook, can be performed using AI and free up humans to be more effective and creative.

The future of AI is probably a lot like the past: it nibbles. Artificial intelligence does a job we weren’t necessarily crazy about doing anyway, it does it quietly, and well, and then we take it for granted. No one complained when their thermostat took over the job of building a fire, opening the grate, opening a window, rebuilding a fire. And no one complained when the computer found 100 flights faster and better than we ever could. But the system doesn’t get tired, it keeps nibbling.

Seth Godin

The New Marketing Landscape

Today, marketers’ biggest challenge is maintaining market share in the saturated digital landscape. If large brands do not consistently advertise and spread their message, they risk losing market share to other large competitors with big budgets or smaller, more nimble startups.

Making matters worse, brands must keep up with dwindling consumer attention and increasing competition. The number of platforms seems to increase daily, while the number of publishers now includes anyone with an account. One-hundred years ago, marketers only had three different media options: print, radio, and direct mail. Television became popular in the 1950s. A few more decades later came the internet, bringing websites, banner advertising, and search. But things took off from there: email, Facebook, Twitter, online video, LinkedIn, Pinterest, Instagram, Vine, Snapchat, and more all came onto the scene within the past 15 years.

Each platform has its own audience and function. Brands have to create unique content tailored to each of those audiences and purposes. Put simply, we went from producing around 10 unique pieces of marketing content per year, to needing 19 different content types, and 14,000 content pieces per year. We have no reason to believe the introduction of new media platforms will slow down. The need for more content will only increase along with it.

It is not scalable or sustainable. At least not without artificial intelligence. With AI, marketers can collect, organize, and analyze data; find insights and automate applying them to future campaigns. But creating enough content is not enough.

Better marketing requires offering value and being relevant to capture and keep consumers’ attention. Industry leaders are adopting a more scientific process to marketing: analyze historical data, make a hypothesis, test, and repeat. While they currently have to wait at least several days campaign data to be analyzed, AI enables real-time metrics that can automatically be tested to improve the campaign immediately, as well as over time. AI can optimize messaging, save time, and scale content, giving marketers free time to ideate new campaigns.

Artificial intelligence for marketing processes data in real time and produces:

Applications of Artificial Intelligence in Marketing

AI for Data Collection and Cleansing

Tracking customer behavior throughout the purchase process is easier than ever. The marketing world has become almost entirely trackable and measurable in the past decade. We have reached the point where every online, and many offline, activities leave a digital footprint. While having all that data is extremely valuable, you need it to be together. In most cases data comes from a variety of sources and formats and resides in silos. Marketers need all the data in one place to connect the dots and make sense of the customer journey.

Making that happen today usually entails compiling data from multiple sources in spreadsheets. Data must be cleansed, standardized, organized, and updated regularly. Technically, AI doesn’t actually do the data collecting – that work is handled by APIs. But it does do the important part of cleansing, consolidating, and, most importantly, analyzing (more on that later). With APIs, machines constantly collect all your marketing data into one place, eliminating silos, and machine learning cleanses it and keeps it current for quick analysis.

AI for Data Analysis

Having big data is not useful unless you can make sense of it. While APIs neatly pull data together while you sleep, artificial intelligence extracts actionable insights to help marketers make better decisions. Rather than days, if not weeks, of human analysis, you instantly have intelligent insights, updated in real-time. Additionally, “[artificial intelligence] is able to process data independently without human biases, manual errors, or hidden agendas. AI is also infinite when it comes to processing data, which means analytics can be scaled up on demand.”

AI systems use machine learning to comb data for patterns between marketing KPIs and inputs. The way machine learning works is by first identifying a goal, then teaching the computer to model conversions by giving it examples of that goal, and letting it continue to improve the model with new data, until it can predict a conversion before it happens.

Think of how Pavlov taught his dog:

  1. He wanted to measure how much saliva dogs produce.
  2. He started ringing a bell when he fed the dog.
  3. Eventually, the dog began to salivate when he heard the bell because it was associated with food.
  4. The output grows better and more accurate with the constant flow of new information.

Of course, marketing objectives are a little more complicated. Fortunately, AI systems can uncover minute patterns in data that would take a person several months, at least, to find. If you want to know what content drives the most engagement on Instagram, you start by connecting your Instagram feed to the system. It will analyze historical post data, breaking down the post into multiple elements and identifying which elements were present in the most engaging posts. As new posts are added, it will continue to process new information until it can tell you what subjects, colors, keywords, publish times, and more will drive the most engagement. The smartest models look beyond just your data, at what is working for your competitors and the industry overall.

Artificial intelligence can tell exactly how your campaigns are performing at any time, and what is causing any increase or decrease in performance. You can think of it like a metrics dashboard, with numbers updating in real time. Only, instead of only showing ad engagement going up, you’ll see why and with which audience. Based on these real-time insights, you can adapt campaigns to improve results.

Using AI to Summarize Data

Many companies, including the Associated Press, Orlando Magic, and Deloitte use artificial intelligence to write stories based on data to produce more content while saving time. It can produce drafts of marketing reports, financial summaries, and sports recaps. AI is quickly becoming the fastest way to translate data into clear summaries.

AI for Consumer Insights

Marketers need to understand buyer behaviors, motivations, and expectations in order to serve relevant messages. Because consumers’ digital trail can be collected and analyzed, we can get much deeper than audience demographics and down into consumer psychographics. AI can look at behavior on a much more granular level and predict what a consumer will do next based on their past actions. It can draw conclusions between certain types of people to anticipate their next move, motivations and desires.

If all the people in one group did A, B, then C, we can predict that other consumers who do both A and B will likely follow with C. This is how computers and advertising platforms create “lookalike” audiences. Users who have previously engaged with a brand, including customer lists, can provide data to help find other consumers with the same behaviors.

AI systems use machine learning to comb data for patterns between marketing KPIs and inputs. The way machine learning works is by first identifying a goal, then teaching the computer to model conversions by giving it examples of that goal, and letting it continue to improve the model with new data, until it can predict a conversion before it happens.

In the past, we looked at demographics because they are a good indicator of general behavior, and that data was easily available. For instance, urban men and women between 18 and 40 years old, with expendable income would be categorized as Starbucks patrons. Dunkin Donuts customers, on the other hand, are generally more suburban, less technically-savvy, and skew older.

Now, we know that psychographics are a much better indicator of motivations, needs, and purchase behavior. For example, two 30-year-old, urban, middle-class white men with identical backgrounds don’t necessarily spend money on the same things. One may prefer to spend on shoes while the other would rather take vacations. Marketers can uncover that information about their target consumers in moments with the help of AI. By looking at audience behavior, AI systems find out the interests, context, hedonistic activities around users and products. And the system automatically adapts with evolving consumer behaviors and interests.

Target Learns Teen is Pregnant Before Her Father Does

After analyzing historical data on customer purchase behavior, Target uncovered a connection between the purchase of unscented lotion and dietary supplements with being pregnant. When shoppers start buying those items, Target sends them offers related to pregnancy and baby items. A man outside of Minneapolis complained to a store manager after his teenage daughter received similar coupons. He later apologized, explaining his daughter was, in fact, pregnant.

AI for Account Based Marketing & Personalized Messaging

Once you understand the motivations behind consumer actions, AI uses those psychographics to develop more personalized marketing messages. Instead of creating a message aimed at a mass market, or even at a very targeted group, AI is leading to markets of one individual. Depending on where someone is within the marketing funnel, and the unique path they took to get there, you can serve them the exact messaging that will drive them to the next step.

Marketers can use this capability to make sure no opportunity is missed or wasted and no lead is left behind. Whether someone is researching for a vacation months away or in the coming weekend, AI can predict his or her needs and keep serving them relevant messaging. A luxury hotel can serve Facebook ads about different activities or amenities for the pleasure traveler, and conveniences for the business traveler. Has this individual stayed at the hotel before? With AI, you can offer dinner reservations at their favorite nearby restaurant or let them know about sales at a local shop they purchased from before. Based on their responses to those messages, AI will create even more personalized messaging in the future.

AI for Programmatic Advertising

Developing relevant messaging is the first step. Next, you want to make sure messages appear in the right context. You wouldn’t try to sell sunscreen when it’s raining or umbrellas when it’s sunny. You also don’t want to advertise a product to someone who has just purchased it.

As in the Target case, brands can use artificial intelligence to serve more relevant, targeted messages in the right places, at the right times. Marketers simply input buyer segments, campaign parameters, such as targets, KPIs (like CTR, ROI, impressions, etc.), and advertising budget and AI finds the appropriate placements for ads. It can adjust placement and spend based on real-time responses. Additionally, if the computer is given different customer segments with varying lifetime values (CLV), AI can adjust bidding strategies and spend more for higher-value target customers. With the adoption of artificial intelligence “marketers are going to gain the opportunity to go back to strategy, content and creative,” letting AI do the hard work.

Cosabella Ditches Marketing Agency for AI

Lingerie brand Cosabella was seeing diminishing returns from their agency’s campaigns, so they decided to consider other options. Instead of hiring a new agency or bringing resources in-house, they decided to use artificial intelligence. After a brief teaching period, where they gave it parameters for targeting and KPIs, they left the AI to itself.

The results were impressive: “After three months with [the AI] Albert, Cosabella saw a 336% increase in return on ad spend. In Q4, revenues increased 155% and the brand saw 1,500 more transactions year over year, 30% of which came from new customers. In Albert’s first month, Cosabella decreased costs by 12% by increasing returns by 50%. Cosabella’s team was able to turn insights into action and spend more time working on more creative strategies.

AI for Customer Communication

With the rise of mobile adoption came the development of mobile marketing and apps. Recently, however, consumers have grown tired of apps, and devote most of the real estate on their smartphones to messaging platforms, like Facebook Messenger, WhatsApp, Slack, and WeChat.

One way for brands to make sure they are a top of mind is with chatbots. Advertising Age defines a chatbot as “an application typically powered by artificial intelligence that is designed to simulate a conversation with another human.”

Chatbots create a more natural, frictionless way of performing an action and interacting with brands. Companies and consumers are already using them in several different ways: to provide relevant content (CNN), make purchasing easy (Taco Bell, Domino’s), and to answer common questions ( “Chatbots in particular sectors do what an app can do but don’t need to be downloaded – they live on servers, not a user’s device – so they are easier and less expensive to create, update and deliver to customers,” explains Christine Duhaime.

In addition, chatbots can facilitate better user experiences in onboarding and customer service. Making sure new customers quickly get to the ‘Aha!’ moment with a platform or find the product they are looking for on a website is critical for retention. Artificial intelligence solutions streamline onboarding by monitoring user activity and prompting them where necessary or offering help. Social media bots can also answer any questions posted by consumers, reducing the response time and making it easier for customer care teams to focus on more complicated cases.

Using a Digital Assistant to Manage Your Calendar and Set Up Meetings

If you have ever received an email from “Amy Ingram,” then you were corresponding with a computer. is an AI personal assistant that integrates with your email and calendar to schedule meetings. It works so well, people think the computer is a real person.

AI for Content Creation

Arguably, the hardest task for artificial intelligence to take over is content creation. Yet, it is also one of the most crippling challenges for marketers. For some reason, the fundamental unit of marketing, a piece of content, is still developed as if creativity is a pursuit immune to numbers.

Few people can write variations on the same message indefinitely without getting bored. But lots of content creation is surprisingly repetitive, and machines have improved in writing to the point where it’s hard to tell the difference. Chances are, you have read articles written by machines without realizing it. Financial summaries, sports reviews, and other quantitative analyses are ideal for automated narratives generated by artificial intelligence. As a matter of fact, artificial intelligence can go beyond turning numbers into written summaries.

Advances in Natural Language Processing (NLP), image recognition, and machine learning have given AI the ability to predict what messages and images will drive desired consumer actions. Each engagement with brand content is another data point teaching machines the content they want to see and how they want to consume it.

Coca-Cola Investing in AI for Content Creation

Coca-Cola’s global senior digital director, Mariano Bosaz, announced the brand would be pursuing ways to implement automated narratives, including writing social media posts, scripts, and even choosing music. The brand recognizes the challenge of scaling creation of smart marketing content. In order to consistently entertain, inform, and offer value to its customers, on each of the platforms they use, Coca-Cola is investing in artificial intelligence to supplement their team.

Where Does AI Fall Short?

Despite all its benefits, artificial intelligence is not the silver bullet to incredible marketing. AI will not take the place of a marketing team because it cannot replace creative thinking. Because it is a program, AI does not have new ideas – only variations on older campaigns.

Furthermore, as with human analysis, you have to ask the right questions (or program the right instructions) to get the answers you’re looking for. An AI solution is only as good as the data it has. Incomplete, inaccurate, or outdated information will result in suboptimal output. Marketers should be careful how and what AI systems are taught, as Microsoft learned the hard way with Taybot. Like toddlers, artificial intelligence solutions are very impressionable.

How to Implement and Leverage AI

Harnessing big data is the first step towards adoption and success with artificial intelligence. All the data exists, marketers just need to make sure they are effectively capturing it. The relationship between marketing and technology departments will become very strong. The best AI solutions will consider your existing legacy systems and allow for collaboration and communication across enterprise.

“Marketers aren’t just thinking about how AI can reduce basic task burdens; they are focused on how it can improve the business, which is exactly what every organization should be focused on with initial investments in AI.”

The terms “artificial intelligence” and “machine learning” are very hyped right now. As you will discover, some AI solutions are better than others. In many cases, humans are still behind the “intelligence.” They should be able to give a stand-alone demo. If they need to process data “in the cloud,” that is a sign they may be relying on human analysts. Ideally, they can give a demo on your data to prove it will actually work with your data. Make sure solutions actually provide real-time results, allowing you to instantly adapt strategy.

Research their data sources and algorithms. As we explained earlier, big data, the more the better, is what makes AI valuable. And delivering a huge data file isn’t artificial intelligence – it’s just a database requiring you to do the work to find insights. Find out what approaches to AI they use. It’s a good idea to have an expert, like a data scientist, with you to understand what they are saying. While you may not understand the technology, the provider should have no problem with talking to you about it.

Make sure it will be able to change along with your needs and grow with your business. Ask the company where they see themselves in five or ten years. They should not only anticipate changes for their products, but also for your business needs.


A well-rounded marketing strategy will incorporate a diverse set of activities and tactics. “If [data, intelligence, execution, and deployment] are the layers of our marketing cake, the reality is AI can’t now be a module, it has to vertically integrated with all of that – it needs the data, the intelligence, the segmentation, then access to deployment engines in order to execute a campaign,” explains Allen Nance, Emsarys CMO. AI acts as a resource for the team to perform their jobs better, faster, and more easily. Ideally, it will seem like an additional team member (of the data analyst variety), who never complains and is constantly working, reliably. It always has an answer anytime you ask.

In the future, marketing teams will devote most of their time to strategizing new campaigns, writing thought-provoking, novel content, and finding new ways to offer value to their customers. Meanwhile, their AI support systems will give direction for optimizing content, predict how messages will perform, and directly tie marketing efforts to business results.

Ready to Optimize Your Social Content with AI?

Get Started!