AI in marketing: How to find the right use cases, people and technology

Posted on

Most marketers already know they can capitalize on artificial intelligence (AI) to make more informed decisions, better engage their target audiences, and drive revenue for their organizations.

Yet, according to a Demandbase survey released in 2019, only 18% of B2B marketers and sales professionals are currently using the tech.

The same study also found that 67% of marketers expect higher lead quality from AI, and 56% believe the technology can help yield better engagement with customers and prospects.

So, what’s holding marketers back from using it?

While marketers recognize the value that the tech can deliver, they often lack the perfect combination of prioritized sweet-spot use cases, people/organizational capacity, and technology to effectively execute an AI strategy.

Unfortunately, by not mastering this trio, marketers are putting themselves—and their companies—at risk of becoming obsolete.

Experts from McKinsey & Company predict that AI technologies could lead to a substantial performance gap between front-runners (who fully absorb artificial intelligence tools across their enterprises) and non-adopters or partial adopters by 2030.

AI front-runners are projected to potentially double cash flow by 2020, with implied net cash-flow growth of roughly 6% for through 2030, while non-adopters “might experience around a 20% decline in cash flow from today’s levels.”

To avoid falling behind and to begin reaping the benefits, every marketer must prioritize identifying the best-fit use cases, hiring and/or developing the right people, and implementing the right technology in the year ahead.

The AI landscape is littered with failed projects, so here’s what to keep in mind to increase your likelihood of success:

Identifying the best-fit AI use cases

While there may be hundreds of AI use cases that a marketer will eventually want to execute on, marketers should first map out their top candidates according to two dimensions: value and feasibility.

It’s okay to first think big, but then you need to narrow the list.

Among the common use cases are the following: intelligent chatbots, smarter personalized digital advertising, content generation and curation, AI-powered account or lead scoring, AI-assisted email responses, multi-channel marketing attribution, next best action, customer lifetime value, and sentiment analysis.

Marketers should estimate the value delivered for each use case (potential upside revenue, time-to-market, reduced manual labor, customer satisfaction), as well as time and effort it will take to see actionable results.

If the use case isn’t both highly valuable and highly feasible – and if you don’t know how you’ll act on the predictive results – then it should be taken off the short-term wish list.

Marketers who are unsure of where to start should consider assessing the value of these common high-impact applications:

  • Optimizing advertising spend: Marketers spend billions of dollars a year on advertising, but often have no way of quantifying whether these investments are worthwhile. With AI, marketers can more accurately attribute sales to specific advertising initiatives, enabling them to optimize their spend to bring in more leads for less resources.
  • Enhancing customer experiences: AI can empower marketers to hone in on their customers’ preferences and create personalized experiences based on past buying and browsing behavior. Not only does this enhance the customers’ perception of the brand, but it can also lead to increased sales—especially when they are recommended a product they hadn’t previously considered.
  • Predicting and mitigating customer churn: Customer retention teams often have limited resources and aren’t able to dedicate the same level of attention to every customer. To solve for this, marketers can implement an AI solution that discovers patterns in historical customer activity to accurately predict which customers are likely to leave them for a competitor. Using this information, the team can better focus retention efforts on the customers that are most at risk and offer them incentives to remain loyal.

Once marketing teams have identified the processes they want to apply AI to, they can start to identify the individuals who will lead the implementations and the technologies they need to bring those use cases to life.

Hiring or developing the right people

The skillsets of the modern-day marketer are fast-evolving.

With the number of digital customer touchpoints that marketers need to manage—which includes everything from desktops and mobile devices, to social media and beyond—marketers need to consume, analyze, and leverage endless amounts of data to inform decisions.

That data is especially crucial for fueling valuable AI applications; without it, the systems won’t have the necessary information they need to generate mission-critical insights—such as predicting consumer behavior or creating truly personalized content.

It’s no surprise then that Marketing Land’s January 2019 Digital Agency Survey found 72% of agency marketers said data science and analysis will be the most in-demand technical skills in the coming years, followed by conversion rate optimization (59%), and computer science/AI and technical SEO (52% each).

Unfortunately, those skills are hard to come by; according to Indeed, the number of individuals searching for AI-related jobs decreased by 14.5% from May 2018 to May 2019. They also found that demand for data scientists increased by 344% from 2013 to 2019, yet the talent pool grew by just 14% in 2018.

Although the talent shortage certainly presents challenges for marketers, there are ways around it. Marketers can identify internal “citizen data scientists.”

These are individuals who possess deep domain knowledge and have a strong analytics background, but not formal data science training.

With the right tools and training, citizen data scientists can get up to speed on the organization’s AI strategy quickly.

Additionally, marketers should consider hiring an AI consultant to support their initiatives or looking to their platform provider for guidance on AI strategies in the near-term while they work on adding AI to their marketing DNA and building it as a competency over the longer-term.

Implementing the right AI technology

Regardless of the use case, there are different approaches marketers can take to leverage AI in marketing processes.

Marketers know well that there are some 7,000+ different vendor tools that could be leveraged in a martech stack, and an exponentially increasing number of those incorporate some AI, or at least claim to do so.

The most common approach taken by marketers today is to leverage AI that comes built-into a martech tool and that is optimized for just that one-point solution or capability.

That means marketers might have 10 different AI tools for ten different capabilities, but that’s the most frequent approach today that gets fast time-to-market without having to hire or develop the AI competency in-house on day one.

While having those point solutions may work today for certain problems, the reality is that some of the highest value problems in marketing or customer loyalty can’t be solved by a point tool.

Use cases such as next best offer, cross-sell/up-sell, churn prediction and reduction, customer experience optimization, price elasticity modeling, customer satisfaction, and others require a broader enterprise solution.

To that end, finding the right AI technology or platform backed by some business transformation help is absolutely critical to marketers’ AI success.

Here are three considerations for success when selecting AI technologies:

  • Automated creation of machine learning models, without requiring coding or data science tools. Not only does this enable non-data scientists to deploy their own models, but it also frees up the experts from the repetitive tasks model building creates, allowing them to use their unique expertise for selecting and fine-tuning models to meet marketing needs. Those steps include preparing the data, modifying it to improve the models, diversifying the algorithms, and more.
  • Monitoring of how models are performing. This is crucial to ensuring the success of the algorithms, as a monitoring component can identify and solve for performance issues, infrastructure challenges, and changes in data. Without the ability to monitor and manage deployments, it’s likely that the AI models will eventually fail.
  • Trusted, explainable AI. Marketers should only invest in an AI tool if it’s human-friendly and the AI can be explainable—in other words, is a “white box” solution. Otherwise, they won’t have any insight into the decisions their algorithm is making and why those decisions are being made. As a result, the algorithm might be inadvertently biased, which could lead to compromised brand reputation and a loss of consumer trust—both of which were top AI bias concerns for the more than 350 U.S. and U.K. executives polled in this recent survey.

The impact of AI is being felt across all industries, and the savviest marketers are prioritizing getting their AI strategies in motion to maintain their organizations’ competitive advantage.

But in the AI-driven era, it’s not enough for marketers to be interested in AI; to be truly successful, they’ll need to think critically about the processes, people, and technology that will be core to their AI missions.

Those that master that combination will be easy to identify, as their organizations will dominate for years to come.

This article was originally posted on –