The AI Marketing Canvas
A Five-Stage Road Map to Implementing Artificial Intelligence in Marketing
Raj Venkatesan and Jim Lecinski



The Marketer’s Challenge Today

EVEN AS IT WAS STRIVING to improve the delivery of its digital products, the iconic news brand The Washington Post was on track to lose an estimated $40 million in 2013, when it was acquired for $250 million in cash by Amazon founder Jeff Bezos.1

In his open letter to its employees on April 5, 2013, Bezos wrote:

There will, of course, be change at The Post over the coming years. That’s essential and would have happened with or without new ownership. The Internet is transforming almost every element of the news business: shortening news cycles, eroding long-reliable revenue sources, and enabling new kinds of competition, some of which bear little or no news-gathering costs. There is no map, and charting a path ahead will not be easy. We will need to invent, which means we will need to experiment. Our touchstone will be readers, understanding what they care about—government, local leaders, restaurant openings, scout troops, businesses, charities, governors, sports—and working backwards from there. I’m excited and optimistic about the opportunity for invention.

Bezos understood that, like so many other businesses, the news business model had shifted from supply to demand because a few big networks (Amazon, Google, Facebook, Netflix, et al.) had reset customer expectations through increasing personalization of the customer experience.

The Washington Post is now a leading technology, software, and media company whose resources allow it to respond to changes in the customer news needs, and to shape those needs by using technology such as artificial intelligence (AI) and machine learning. AI and machine learning allows the Post to market better by personalizing the user’s experience, supercharging it at every juncture. To wit:

  • Search engine optimization serves up relevant stories based on search criteria to potential subscribers.
  • Readers who come directly to the site can experience a limited number of articles for free, enabling them to experience the quality of the journalism while providing the Post with the opportunity to stimulate them to subscribe.
  • New subscribers and casual readers alike continue to experience the quality of the product, from load speed to personalized content driven by AI and machine learning through a system called Zeus Insights, which serves up stories based on interests inferred from previously read articles. AI and machine learning technology also allows the Post to rapidly predict the popularity of articles, so that the newsroom can add media and links to those articles that are trending.
  • Subscribers also are encouraged to add their comments to articles, in effect becoming part of the editorial process. An AI-driven comment moderation system helps the Post to maintain a high-quality comment section, the best of which is rolled into an adjunct publication composed of the best, most relevant comments.
  • The Post’s AI-powered story-writing program called “Heliograf” was instrumental in its ability to produce twice the number of stories as the New York Times (500, versus 230 in 2017).2

The Post’s success is evident: its digital subscriber base has more than tripled in the last three years, and added well over a million new and exclusively digital subscribers in that time frame, according to a May 16, 2019, press release.3 But that’s not all. The Post’s main technology platform, Arc Publishing, is so successful that the Post has licensed it to other top publishers, broadcasters, and brands.4 The technology has become a new business line.5 All of this has returned the Post to profitability for at least two years running.6

If your brand doesn’t have the resources of a giant network such as Amazon at its disposal or a technology visionary with deep pockets at the helm of your company, don’t despair. There are significant gains to be had by implementing AI and machine-learning technology to supercharge your brand’s customer journey by delivering the personalization that consumers now expect, no matter where you’re starting.

Consider CarMax, America’s largest used car retailer, where the majority of auto sales still occur in a physical store, but whose customers’ buying journey increasingly begins online. personalizes the images shown on its website based on your search behaviors. In fact, the text and images change depending on what the site learns about you as you explore the site.’s goal is to present you with increasingly relevant and desirable inventory, thus reducing the cognitive load and removing friction from your research process. This differentiates from its competitors because the site is explicitly designed to reduce the anxiety that occurs when you are researching a purchase and are overwhelmed with too many choices, particularly when it comes to high-cost items.

According to Barry Schwarz, author of a book on consumer anxiety titled The Paradox of Choice,7 says that the more choices we have, the less satisfied we become. In an interview for about car buying, Schwarz said, “If you buy the wrong cereal, you get to correct that mistake next week. Large decisions are not easily reversible. That’s why there is extra anxiety baked into them.”8

CarMax understands that its long-term competitive advantage hinges on its ability to collect first-party data and use it to ease the purchase process by personalizing the consumer’s experience—online and everywhere. CarMax also realizes it is not competing against the best experience consumers have ever had buying a car; it is competing against the best experience consumers have ever had, period.

Let’s take coffee as an example. When you launch your Starbucks app, not only can you order exactly what you want when you want it (e.g., nonfat grande iced coffee with two pumps of toffee nut syrup at 2 p.m.);9 the app also will recommend new drinks and food based on what it knows about your purchase history and preferences. CarMax knows you expect it to deliver this same level of personalization—from the mobile marketing messages to your in-store experience. The car retailer wants to learn as much as possible about you, so that it can anticipate your needs and priorities; and it has invested heavily in technology to be able to do just that.10 (We’ll talk more about CarMax and Starbucks later in this book.)

Other brands are working through the sometimes painful process of transitioning from the old supply-driven or “analog” business model to the on-demand or “digital” business model. Many brands have yet to begin, or are in the midst of building, digital foundations that will allow them to collect the consumer data required for their marketing, and to benefit from the more advanced technology. Even brands that have substantially built their digital foundation may still be in data collection mode, and may thus be focused on learning some lessons and getting some quick wins, leading to insights that are cumulative and can be built upon.

For example, the first generation of Coca-Cola’s Freestyle dispenser allowed customers to personalize their soda by mixing and matching different flavors, and this generated a ton of data about consumer preferences. The machines sent the data about combinations customers created all around the world back to corporate headquarters. This digital initiative led to the development of a successful new product in Sprite Cherry, and to an even more sophisticated dispenser that does provide one-to-one personalization: the Powerade Power Station.

On the other hand, the cost of delaying or ignoring the need to make AI and personalization a key strategic objective can be steep. For example, Kraft Heinz has signaled its intention to incorporate AI and machine learning into its operations after experiencing a sales slide and significant write-downs in the value of some of its most prominent brands in February 2019—turbulence believed by analysts to be the result of management’s previously rigid philosophy of growth through cost cutting. The company appointed a new CEO, Miguel Patricio, who was formerly CMO at Anheuser Busch InBev;11 and less than six months later it announced a new CIO in Corrado Azzarita. According to, Azzarita said he intends to implement “machine-learning models that crunch data such as historical sales, rivals’ current promotions, and macroeconomic variables to recommend optimal promotions for Kraft Heinz brands, and other models that help it figure out the best mix of media to use to advertise products.”12

Still, a recent study of three hundred advertisers by Advertiser Perceptions found that half of the marketers surveyed have no plans to use AI in their marketing. Said Frank Papsadore, EVP at Advertiser Perceptions, “Big-budget brands like Nike, IKEA. and Sephora are pioneering AI for marketing, but most advertisers don’t have their resources, so they’re focusing on more immediate marketing efforts.”13 This means if you are a marketer at a well-established firm, you may face some major internal barriers to the process of implementing AI and machine learning, some of which may require what will be rightly perceived as radical changes.14

The problem with this hesitation to move forward is this: Everywhere across the business landscape, the high-margin analog business model is on its way out and is being replaced by a lower-margin, digitally driven business that relies on volume. The significant investments required to create a successful digital business could mean that profits will decline before the brand emerges on the other side to new, lower unit profitability, but higher volume and lower costs, and potentially higher total profits.15 To succeed, management must make a commitment to investing long term and be prepared to tolerate a temporary chasm of low or no profits. Otherwise, their brands may face bankruptcy—and they may not come out alive.

Consider all of the brands filing for Chapter 11 bankruptcy: Sears, Claire’s, Toys “R” Us, to name but a few.16 Sears, for example, has shrunk its physical footprint by 75 percent, sold critical assets, and laid off thousands of workers from its corporate offices and stores. It has emerged from bankruptcy and is opening new stores for home goods.17 Claire’s Stores, the US retailer popular with teens as a destination for ear piercing, affordable jewelry, and fashion accessories, also recently filed for bankruptcy. In doing so, it closed stores, removed $1.9 billion of debt, gained access to $575 million in new capital, and announced plans to reinvent itself as a smaller, more profitable business.18

Toys “R” Us shuttered its stores in 2018, with networks such as Amazon swooping in to fill the void. It was purchased in 2019 by Tru Kids Brands.19 Tru Kids plans to revive the brand through opening smaller “experiential” stores, and is collaborating with a variety of other retailers including Target. It is also partnering with the interactive candy experience purveyor Canditopia,20 to create Toys R Us Adventure, a series of interactive playrooms featuring installations that put a spotlight on Geoffrey the Giraffe, the Toys “R” Us mascot. Tru Kids Brands has partnered with the retail-as-service startup b8ta, who will be giving Tru Kids access to data and analytics to track things like foot traffic in and out of the stores to allow the company to make smarter decisions, according to Phillip Raub, b8ta cofounder and president, who says, “This year [2019)] is going to be an opportunity for us to test and learn.”21 If the brand wants to use data to promote the experience online, however, the increased enforcement of the Children’s Online Privacy Protection Act (COPPA) needs to be considered to develop a responsible and ethical online strategy.

The point is, it’s possible to survive bankruptcy—as well as other crises such as a pandemic or stock market bubble—and to live another day. But to survive in the longer term, we believe there’s no avoiding this transition. Brands have to figure out how to leverage AI to serve consumers better. Why? Because the resources required to respond to this new demand economy—the demand for one-to-one personalized marketing—far outstrips human (analog) capacity.

While there’s still time to catch up, there’s no time to lose. Time is of the essence because, unlike the technology advances of the late 1990s, the advances experienced by brands that commit to applying AI and machine learning to their marketing now will be exponential. Insights obtained from customer data are cumulative, and the custom AI algorithms that are developed to leverage these insights will allow firms to provide personalized experiences—from advertising to pricing to point of sale of promotions and beyond—whose value for customers will increase exponentially over time.

In this new AI-driven business model, the winner takes all. Those who wait to apply AI and machine learning to their marketing will be left behind—and that could mean the end.

The good news is that there is still time to get in the game—if you start now. This book aims to show you how.


1. Jeff Bezos, “Jeff Bezos on Post Purchase,” Washington Post, August 5, 2013, accessed June 15, 2020, 2013/08/05/e5b293de-fe0d-11e2–9711-3708310f6f4d_story.html.

2. Claire Atkinson, “The Washington Post Still Plays Catch-Up, but Is Gaining on The Times,” NBC News, December 28, 2017, accessed November 1, 2019, washington-post-still-plays-catch-gaining-times-n833236.

3. “Arc Publishing Unveils State-of-the-Art Commerce Platform Arc Subscriptions,” Washington Post, May 16, 2019, accessed November 4, 2019, unveils-state-of-the-art-commerce-platform-arc-subscriptions/.

4. “Arc Publishing Licenses Technology to Global Brand BP,” Washington Post, September 25, 2019, accessed May 22, 2020, arc-publishing-licenses-technology-global-brand-bp/.

5. “NPR: At ‘Washington Post,’ Tech Is Increasingly Boosting Financial Performance,” Washington Post, June 14, 2017, accessed November 1, 2019, npr-at-washington-post-tech-is-increasingly-boosting-financial-performance/.

6. Sara Fischer, “Scoop: WaPo Hits 2nd Year of Profitability, Plans Expansion,” Axios, January 9, 2018, accessed November 1, 2019, -4e99-b002-ad41416737ef.html.

7. Barry Schwartz, The Paradox of Choice: Why More Is Less (New York: Ecco, 2004), 2.

8. Rick Press, “Wonder Why You Dread Car Buying? A Famous Psychologist Explains,” Capital One, March 28, 2018, accessed January 23, 2020, wonder-why-you-dread-car-buying-a-famous-psychologist-explains/1030.

9. Lee Breslouer, “Starbucks Baristas Name Their Favorite Drinks,” Thrillist, November 10, 2016, accessed January 23, 2020, drinks-according-to-baristas-who-serve-them.

10. Jim Lyski, “Don’t Settle for the Best Customer Experience in Your Industry, Deliver the Best One—Period,” Think with Google, October 2017, accessed November 4, 2019, experience-design/carmax-industry-consumer-experience/.

11. Lauren Hirsch, “After Brutal Year, Kraft Heinz Taps AB InBev’s Miguel Patricio to replace CEO Bernardo Hees,” CNBC, April 22, 2019, accessed January 23, 2020, kraft-heinz-taps-new-ceo-ab-inbevs-miguel-patricio.html.

12. Martin Giles, “Kraft Heinz Appoints New CIO to Deliver an AI Growth Recipe,” Forbes, November 14, 2019, accessed May 22, 2020, kraft-heinz-cio-uses-ai-machine-learning/#148c96fc28f6.

13. “Contrary to Hype, Advertisers Divided on AI,” Advertiser Perceptions, March 3, 2020, accessed March 4, 2020, advertisers-divided-on-ai/.

14. Lynne Galia and Lainie McKeague, “Kraft Heinz Rebrands Kraft Recipes Website as ‘My Food and Family,’ Adds New Features,” Kraft Heinz, March 21, 2019, accessed October 31, 2019, kraft-heinz-rebrands-kraft-recipes-website-my-food-and-fami-lytm; Peter Eavis, “Kraft Tests How Much Costs Can Be Cut as Tastes Change,” New York Times, February 22, 2019, accessed October 31, 2019, /3g-capital-buffett-kraft-heinz.html; Martin Giles, “Kraft Heinz Appoints New CIO to Deliver an AI Growth Recipe,” Forbes, November 14, 2019, accessed January 23, 2020, /11/14/kraft-heinz-cio-uses-ai-machine-learning/#4d4f0e9e28f6.

15. Sunil Gupta, Driving Digital Strategy: A Guide to Reimagining Your Business (Boston: Harvard Business Review Press, 2018), 192–94.

16. Ben Unglesbee, Cara Salpini, and Kaarin Vembar, “The Running List of 2018 Retail Bankruptcies,” Retail Dive, November 21, 2018, accessed October 31, 2019, the-running-list-of-2018-retail-bankruptcies/516864/;. Daphne Howland, Ben Unglesbee, Cara Salpini, Kaarin Vembar, and Caroline Jansen, “The Running List of 2019 Bankruptcy Victims,” Retail Dive, October 23, 2019, accessed October 31, 2019, -of-2019-bankruptcy-victims/545774/; Ben Unglesbee, “One Year Later: Toys R Us’ Fatal Journey through Chapter 11,” Retail Dive, September 18, 2018, accessed October 31, 2019, later-toys-r-us-fatal-journey-through-chapter-11/532079/.

17. Ben Unglesbee, “Sears Filed for Chapter 11 with Plans to Close 142 Stores—Now What?” Retail Dive, October 15, 2018, accessed October 31, 2019, -with-plans-to-close-142-stores-now-what/539654/.

18. Emily Price, “Teen Discount Jewelry Favorite Claire’s Emerges from Chapter 11 Bankruptcy,” Fortune, October 15, 2018, accessed January 23, 2020, -emerges-from-chapter-11-bankruptcy/.

19. Matthew Townsend and Joe Deaux, “Toys ‘R’ Us, Back from the Dead, Will Open U.S. Stores in 2019,” Bloomberg, June 21, 2019, accessed January 23, 2020, toys-r-us-back-from-the-dead-will-open-u-s-stores-in-2019.

20. “Tru Kids Teams with Candytopia for Toys R Us Adventure,” PYMNTS, September 19, 2019, accessed January 23, 2020, tru-kids-teams-with-candytopia-for-toys-r-us-adventure/.

21. Lauren Hirsch and Lauren Thomas, “Life after Liquidation: Toys R Us Stores Will Be Back This Holiday Season, This Time with a Tech Partner,” CNBC, July 18, 2019, accessed January 23, 2020, plots-comeback-with-smaller-stores-in-partnership-with-b8ta.html.