Artificial Intelligence

As marketers attempt to harness a myriad of easily accessible customer information, the influx of data has made their jobs more complex. For those creating video content to reach consumers, artificial intelligence tools hold promise as a way of analyzing the wave of data to quickly make content decisions as well as save time and money, experts say.

Many marketers have AI-based hyper-personalization on their minds, but only 9% of marketers have used these tools, according to Ascends2's Hyper-Personalization Strategies Survey Summary Report. Video is already essential to marketers, because the medium delivers high conversion rates and builds strong connections with consumers. The question marketers have to answer going forward is how to get the most insights out of the data they obtain. Most businesses aren't doing well on this front so far, according to interviews with several experts, and may welcome the help of AI, although it comes with its own set of challenges.

"A lot of the hype is, 'buy the AI and forget about your problems.' And I've read a number of vendors' statements that this makes it easy, and that simply isn't the case," Mitch Ratcliffe, partner at Metaforce, told Marketing Dive.

Disappearing barriers

For marketers and advertisers, a big attraction for using artificial intelligence tools to create video content is the potential to save time and money, according to Tim Parkin, president of Parkin Consulting. The real-time feedback offered by AI tools allows marketers to shave off the time it previously took to find out which ads or branded content work in which market, he noted.

Gavin Twigger, chief creative officer of Ansira, noted that AI video software also saves time by automating content creation processes such as color correcting across different frames, audio level monitoring and frame selection.

"It's almost come full circle for me where some of the things I wanted to do years and years ago, 12, 13, 14 years ago when we were working in the direct-to-consumer space, we were limited by the speed at which we could get the information back," Twigger told Marketing Dive. "Those barriers are gone now."

Advertisers can also tap AI tools to make different versions of videos with varying special effects and transitions. Marketers want to be able to compare which clips perform well at the beginning or the end of Facebook video ads as well as which transition effects perform better on the platform across different market segments, Parkin reported. Using AI tools, advertisers can conduct these experiments at scale without asking their video team to spend hours testing out different videos.

"[The payoff comes] once the AI gets to a point where it can help you discover what's working and can get that feedback loop closed in terms of, 'Okay, I produced a video. Tell me which ones were good and which parts of it were watched,'" Parkin explained.

Extending investments

For companies with a library of content, AI tools can also provide a competitive advantage, because they can create new video content quickly for various platforms without having to send out the same videos too frequently — something that can be a turn-off for consumers — Parkin continued.

AI enables advertisers to pinpoint where customers stop watching videos and what information was discussed before they stopped viewing the video, allowing them to recut ads almost in real time based on how viewers interacted with it, Metaforce's Ratcliffe told Marketing Dive.

"By taking your assets and starting to recombine them, you extend the investment you made in content production, because you're reusing it in so many different ways," Ratcliffe said.

In a follow-up email, Ratcliffe pointed to a network marketing company he co-founded that used AI to equip newly hired sales representatives with custom content to share with prospects. As the sales representatives sent content to potential customers, the client used AI to see when changes in wording in the audio and video increased customer response rates.

"Within a month of launch, the client's sales success rate within the first two weeks of a rep's career with the company had increased 60% over pre-launch," Ratcliffe wrote. "That is, 60% more of its salespeople closed a first sale within two weeks of joining the company."

Still just a tool

Although AI tools can reconfigure video content in different ways, marketers and advertisers must continue to create quality content and have an extensive content library, according to Parkin. AI tools may be able to develop variations of the same video, but it's important not to "throw garbage into a tool and rely too much on the tool to turn it into gold," he said, adding that, when handled correctly, machine learning can tell companies where they lack content that leads customers toward purchasing information or other useful material.

As marketers measure the effectiveness of branded content and ads, AI and machine learning distill massive amounts of information to inform them how and which audiences to target next, Twigger said.

Marketers can also hyper-personalize their content to target audiences based on data points like geography, weather or traffic. Geographic targeting is particularly useful for franchised brands, Twigger noted, because marketers and advertisers have to craft content that resonates with local consumers while adhering to the umbrella company’s messaging.

"Taking the brand message and flavoring it for a local audience using all of those inputs — you can't do that on a piece of paper," Twigger said.

Overstepping with hyper-personalization can come across as creepy for consumers, especially as concerns around data privacy continue to grow. With this in mind, Ratcliffe suggested marketers use AI and machine learning tools to not simply persuade consumers but to listen to what they want. Companies that don't respect consumers' privacy and rely on these tools as a way to control consumers could see a pushback, he added.

"We're on the verge of an explosion of storytelling in society," Ratcliffe said. "If we found a way to be privacy respectful so that we weren't strip mining one another of our personal data in order to sell that data … we'd be in a much more trusting and potentially transparent marketplace."

JPMorgan Chase inked a five-year, enterprise-wide deal with Persado in July 2019 as the financial services giant turns to machines to handle more of its marketing duties, according to details shared with Marketing Dive. The company will leverage Persado's artificial intelligence (AI) technology to generate copy for direct response marketing campaigns across its personal banking, home lending and wealth management segments, and for its digital advertising.

The two companies first partnered in 2016, when JPMorgan Chase used Persado's Message Machine database to tag words and phrases that would help it create more resonant copy around Chase's Card and Mortgage business. Their pilot showed that ads with copy rendered by Persado received as much as a 450% lift in click-through rates (CTRs) versus other ads written by humans that commanded a 50-200% lift.

JPMorgan Chase is making a bigger bet on automating key parts of the marketing process with the five-year Persado deal. Marketers' investments in AI and related fields like machine learning have steadily risen along with their responsibilities to master forward-looking technology and drive growth for brands. Many applications of AI are centered on improving personalization and campaign performance. Concerns that AI could potentially sap the human touch from messages might be quelled by Chase's tests of the Persado solution, which showed higher click-through rates than traditionally generated copy.

"Machine learning is the path to more humanity in marketing," Kristin Lemkau, JPMorgan Chase's CMO, said in a press statement. "Persado's technology is incredibly promising. It rewrote copy and headlines that a marketer, using subjective judgment and their experience, likely wouldn't have. And they worked."

JPMorgan Chase applying Persado's software across its direct response and digital display channels signals confidence that the technology will have the capabilities to meet the needs of millions of customers, and eventually provide personalized messages on a more individualized basis. The company hopes to use the solution to improve internal communications and customer service functions as well, according to Lemkau.

Other brands in categories like retail and hospitality have adopted a similar approach. Macy's in September 2018 expanded a relationship with the agency Madras to automate a high volume of creative assets, including email and site content, for its e-commerce platform. Persado also helped Caesars Entertainment leverage AI in email marketing copy to drive a 12% increase in open rates and a 24% increase in CTRs.

A quarter of companies reported using AI at scale in their personalization efforts, viewing it as a crucial tool, according to a survey by Arm Treasure Data and Forbes that was published in spring 2019. The research revealed many still struggle with execution, however, with 59% of respondents reporting that serious personalization efforts are still a year away from being realized.

Consumers are developing an appetite for customized digital content and products made with them in mind. However, marketers are finding the data and tech tools required for this one-on-one service can be challenging to harness. There is also an ongoing discussion around data privacy and how much personal information most consumers will ultimately be willing to share for a customized product or loyalty program. This area could be a stumbling block for marketers, with a survey by ExpressVPN finding that a majority of U.S. consumers are concerned about how brands garner and utilize their personal data.

Still, the profit motivation and fear of losing out to a more progressive competitor continues to push development for personalization programs. Another study found that 93% of companies that had a personalization strategy increased their revenue last year. With driving results in mind, more marketers have been evaluating how they can utilize AI to develop customized digital content. The belief is it will save time and money, because for one thing you can get real time feedback on what is working and what it not, and quickly modify content.

Personalization can boost profitability, but a majority of companies are still struggling with the execution, according to a new study by Arm Treasure Data and Forbes that was provided to Marketing Dive. "The Clear Path to Personalization" study found that only 21% of enterprises described their personalization efforts as "highly successful" and had dedicated more than 10% of marketing budgets to it. Among this leading group, 54% said they surpassed their sales goals in the past year. Overall however, 59% of respondents said serious efforts on personalization are still a year away, with 41% seeing it as a high priority now.

A healthy 40% indicated their efforts to personalize have helped maximize sales, basket size and profits in direct-to-consumer channels such as ecommerce. Moreover 25% of companies said they are now utilizing artificial intelligence at scale in their personalization efforts and see it as a crucial tool to do so.

Respondents indicated that loyalty programs, mobile apps and product development are the areas that matter the most when it comes to customer personalization. Retailers were singled out as a leader in personalization, with 56% indicating a high commitment level, followed by 48% of consumer electronics, 38% of consumer packaged goods companies and 32% of media outlets.

Despite the promise of rewards, there is still a learning curve ahead for personalization strategies. Core issues facing marketers include data quality and data access, as well as change management, the Arm Treasure Data study pointed out. To gather the data, the researchers surveyed 200 marketing leaders and questioned them on their priorities, budget, programs, and how they measure their efforts

Bank of America's AI-powered assistant Erica reached 7 million users in the 12 months following its June 2018 launch in a strong sign of consumer acceptance. The bank has more than 37 million active users of its digital banking services, including more than 27 million mobile users, giving the company a large audience for introducing new services and tech like digital wallets, voice features and loyalty programs.

Erica completed more than 50 million client requests using artificial intelligence (AI), predictive analytics and natural language processing (NLP). Those technologies have doubled Erica's conversational knowledge to more than 400,000 ways to ask financial questions from 200,000 at launch. Since expanding Erica's features last year to show customers a broader view of their finances, an average of 150,000 customers a week now use the service.

Erica's significant usage demonstrates how consumers are growing more comfortable interacting with virtual assistants for some tasks while signaling that AI and natural language processing can give mobile apps a wider range of functionality for on-the-go customers.

Here, the national bank is flexing its digital tech muscles as mobile banking continues to grow more popular, with more than 53% of retail banking customers using mobile financial services, according to market researcher J.D. Power. 

Nearly half (49%) of Erica's users are millennials, followed by Generation X (20%), Baby Boomers and seniors (16%) and Generation Z (15%). The number of users are likely to rise as tech-savvy Generation Z continues to reach adulthood and begins using more financial services.

While banks have a worse reputation now that they did before the financial crisis of 2009, only 4% of consumers switched banks in the past year amid improvements in customer service, per J.D. Power. Personalized service from AI-powered tools like Erica will be a key part of boosting customer trust in retail banks, especially among younger adults who are more familiar with digitized financial services. Banking apps are best at helping customers transfer funds, pay bills and perform routine services, but are more limited in helping people manage their finances and reassuring users about security and privacy, per Forrester Research.

Walmart brought AI out into the open in April 2019, as a Walmart Neighborhood Market in Levittown, New York was transformed into the company's first Intelligent Retail Lab (IRL). The 50,000 square-foot store is equipped with AI technology to enhance the shopping experience and improve business operations.

The IRL store also featues AI-enabled cameras, sensors, interactive displays and a 100-server data center that is visible to customers. Hardware in the store is connected with enough cabling to scale Mt. Everest five times and enough processing power to download 27,000 hours of music per second, according to the company. 

Walmart will use AI throughout the store to gather information, track inventory, determine whether or not meats and produce are fresh and identify spills or hazards that need attention. AI tools will also help Walmart reduce lines and wait times in store, replenish shopping carts and ensure online orders are always ready for in-store pickup.

Although AI is becoming increasingly important to grocers, it's an amorphous concept for most consumers. Removing the mystery of in-store technology could be key in gaining shopper trust and helping them understand how AI can work in beneficial ways.

As soon as shoppers walk into the IRL store, they see bright colors, techie decor and a glass-encased data center illuminated for customer viewing. The space also offers educational kiosks upfront where customers can learn more about the technology, and an interactive wall shows shoppers how AI can track body positioning.

"We chose right from the very start to not hide the technology," said Mike Hanrahan, CEO of IRL, in a statement.

With its many cameras looking down from the ceiling, the IRL store brings to mind Amazon Go. But where that concept centers on frictionless checkout, Walmart's sensor network primarily monitors shelf performance. Walmart has also rolled out aisle-scanning robots in more than 300 stores as it seeks to improve store efficiency and get the most out of its workers.

Walmart said its Levittown Neighborhood Market on Long Island is one of the company's busiest banner locations, and if the IRL concept works well there and resonates with shoppers, the retailer is confident it can be rolled out on a broader scale.

For now, the company is focused on starting with real-life, practical applications for AI that will make the most difference to shoppers and Walmart associates on a daily basis. From tracking the freshness of meat and produce to opening more cash registers when lines get long, customers are likely to see a more efficient shopping experience and employees will be freed up to focus on customer service.

"Technology enables us to understand so much more — in real time — about our business," Hanrahan said. "When you combine all the information we're gathering in IRL with Walmart's 50-plus years of expertise in running stores, you can create really powerful experiences that improve the lives of both our customers and associates."

Brick-and-mortar retailers are quickly adopting AI to improve efficiency and streamline operations. Ahold Delhaize recently introduced an AI supply chain solution to help manage its perishable offerings, which will be rolled out across its banners in the next three years. Giant Foods has deployed "Marty" the robot to 172 stores, with the goal to expand its functions as the banner better understands how to apply AI. Kroger has plans to open as many as 20 automated Ocado fulfillment centers, which will use AI to speed up online grocery order fulfillment. The first facility is under construction in Monroe, Ohio.

L'Oréal's Modiface unit launched a skin diagnostic tool in February 2019 that's powered by artificial intelligence (AI), letting people upload selfies to get treatment advice, according to a company press release. Development of the digital tool is based on 15 years of scientific research on aging skin by the company's research and innovation teams. The tech leverages an AI-powered algorithm developed by Modiface that was trained on 6,000 clinical images from L'Oréal's research, and on more than 4,500 smartphone selfies of women with different racial backgrounds in various lighting conditions, per the press release.

The new tool serves as the foundation of the company's Vichy SkinConsultAI offering, which launched in January in Canada and will expand to other L'Oréal brand sites around the world through 2019.

The beauty giant acquired Modiface in March 2018 after the startup had a productive few years launching beauty tools that used augmented reality (AR) through Facebook Messenger, as well as working with brands like Smashbox on innovations such as a virtual makeup try-on app with eye-tracking capability. Several months before the acquisition, Modiface developed its own skin assessment capability that was tested and adopted by L'Oréal's Vichy skincare brand. The newer diagnostic tech is an evolution of that initial offering.

"With the acquisition of ModiFace, we have started a second phase of L'Oréal's digital transformation, focused on reinventing the beauty experience through technologies such as voice, AR and AI," Lubomira Rochet, L'Oréal's chief digital officer, said in a statement. "We believe that services will be the new gateways for discovering our brands and products."

Global spending on AI will reach $35.8 billion in 2019, a 44% increase over 2018, according to forecasts from the International Data Corporation (IDC). By 2022, spending on AI systems is expected to more than double to $79.2 billion, with a compound annual growth rate of 38% from 2018 to 2022. The U.S. will account for nearly two-thirds of all AI spending in 2019.

Retail has been leading investments in AI systems, and was projected to invest $5.9 billion in 2019 on solutions like automated customer service agents, shopping advisers and product recommendation platforms. The sector is topping AI spending as personalization becomes a more important driver of e-commerce sales and consumers demand better customer experiences, including those delivered by automated agents like chatbots and business messengers. 

In fact, automated customer service is expected to see the biggest investment this year acros categories at $4.5 billion, followed by sales process recommendation and automation at $2.7 billion and automated threat intelligence and prevention at $2.7 billion, per IDC. Automated prevention maintenance, diagnosis and treatment systems, fraud analysis and investigation, intelligent process automation and program advisers and recommendation systems are expected to see spending exceed $2 billion in 2019. 

Several recent case studies highlight how retail marketers are deploying AI to support their business goals. The men's apparel brand DXL, for example, used the AI-powered interactive content solution InHabit for a campaign that delivered an engagement rate of 6.5% and a dwell time of 21.6 seconds. Six percent of users who engaged with an InHabit unit clicked through to the DXL website, and more than 90% of InHabit engagements were from DXL's target audience of men in the U.S. aged 18 to 50.

Along with personalization and customization, marketers in retail and elsewhere also see AI as helping them boost their campaign effectiveness and enforce security. Banking is expected to invest $5.6 billion in AI, including automated threat intelligence and prevention systems, fraud analysis and investigation. Discrete manufacturing, health care providers and process manufacturing round out the top five AI sectors, IDC said. 

Mounting interest in AI technology for marketing purposes extends to the enterprise sector as well. More than 30% of surveyed B2B marketers believe AI will offer more detailed analysis of campaigns, and 59% report the technology will help them better identify prospective customers, according to EverString and Heinz Marketing research.

While marketers see the benefit of AI, just 9% have a hyper-personalized strategy, which refers to marketing that leverages AI and real-time data, Ascend2 found in a recent study. Most marketers surveyed, or 62%, are just talking about hyper-personalization or haven't done anything concrete about it yet. Improving customer experience and applying data insights to decision-making were top priorities for those concerned with realizing hyper-personalization.