Mark Tarro, Senior Solutions Engineer, Oracle
One of the first AI case studies most people are familiar with is the Gary Kasparov vs Deep Blue chess match. Kasparov, one the world’s greatest chess players, was bested by artificial intelligence in 1997 and the world began to realize it would be “impossible for humans to compete” with technology advancing, combined with humanity’s propensity for error.
AI’s Migration into B2B Marketing
AI has crept into everyday life in a number of ways, such Alexa and Siri. When Spotify recommends new songs or artists to me – it has all the necessary data to make good decisions. When two opponents engage in a chess match, the piece placement and relativity to other pieces are data points the players can utilize to make the right decisions and moves.
However, AI’s slow adoption within companies is mostly due to disconnected customer data across an organization, making it hard to drive appropriate AI decisions. When marketers work with fragmented data, that limited visibility leads to broken customer journeys and experiences.
With marketing, we’re making decisions to guide our customers and prospects along a journey. These decisions include; who to include within campaigns, which way to route contacts within those campaigns, how to score contacts based on their interactions, and which product or service is best to offer. This is where aggregating and connecting data will be key as B2B marketers increase their dependence on AI. Understanding that prospects are likely to purchase based on certain attributes can help with predictive scoring. The more data you connect on the customer profile, the more likely those predictive scores will be effective. The same can be said for predictive content, segments, or any time of an AI-assisted decision your team makes.
Bigger is Better
To understand why a bigger pool of data is better, I started out with my typical Michael Scott quest to find someone to “explain it to me like I’m five.” What I came up with revolved around Machine Learning and pumping them full of data to learn. When I was a graduate student, I worked on a project where I manually assigned numbers to keywords from interviews and counted how many times those keywords were used. As part of the academic study, this project would show valuable learnings from the interviews. However, it was a tedious process that included hours in front of a computer screen. The propensity for human error that Kasparov mentioned is sure to have caused some imperfections in that study, in addition to the weeks it took me to complete. Had a machine been taught to work through that process, it would have done so more quickly and without error.
Machines can also take it a level deeper with Deep Learning. With my example, the data was all connected and related in a straightforward way. More and more of what we’re seeing is companies collecting data from a multitude of channels, sources and devices, but not knowing what to do with it. Valuable data is often discarded when it could be connected and mined for deep insights to inform decisions. Imagine knowing that your chess opponent wearing a certain brand of watch tips their strategy and moves. This piece of seemingly unrelated and unimportant data could be most valuable when looked at in a larger pool. B2B marketers who are active in the social media space might be familiar with Latent Semantic Analysis or Natural Language Processing which look at things like sentiment and possible actions based on social posts.
These technologies are examples of deep learning in action. Where people might take months or years, AI finds connections in real time, reducing human error and identifying subtle nuances, like not letting a tweet intended for customer service slip through the cracks because you hadn’t had coffee yet.
Working with Sales
Predictive lead scoring can point salespeople in the right direction as they work to expand within a key account and understand whom the right contact will be to engage with next. As they email the prospect and drive them to a website utilizing AI, personalized and predictive content can be served up to enhance the experience. This experience can only be optimized based on properly connected data so the predictive scores and content are aligned with marketing efforts. Imagine trying to predict content for someone based solely on his or her job title! You might see some success, but not as much if you were also taking into account their location, company size and any other relevant data points.
As we stare down the inevitable partnership humans will have with intelligent machines, it will be key to figure out which use cases make sense for AI, and which don’t. Most times the decision will come down to having the necessary data being available and connected to be effective. Investing time and effort now into connecting data will allow you to work more harmoniously with AI solutions and apps as they hit the market. Work with your sales counterparts to vet your strategy. See where AI fits and then, integrating your strategy with AI efforts will allow you to compete with the machines instead of against them.
I recently spoke with Jeff Wartgow, Oracle’s Senior Director of Product Management for Fusion CX Strategy. Here is part...