Hey Siri:

“Write a blog about our latest brand mentions in Shenzhen, China and Bogata, Columbia to include sentiment, peak times when our brand was mentioned, and data on peak purchasing times with product SKUs and images. Thanks Siri. Oh, and another thing, please update our landing pages to reflect this data in the native language of the region. Thanks again, Siri.”

When Doug Dawson wrote his article in February on Artificial Intelligence, he felt that #ai is saddled with too much hype as is its current counterpart, 5G. Certainly, there is a great deal of technology out there that deserves the hype and others that do not. After receiving a certification (Intro to AI) from IBM (while using Watson), I’m here to share that artificial intelligence, and its subsets of machine learning and deep learning definitely deserve a soundbite.

While there are many tentacles to AI, including but not limited to, machine learning and deep learning, connectors, and data-lakes, I will attempt to underscore why the impetus to utilize unstructured data is the most important data set for today’s digital marketing initiatives.

Forget your structured data and your hell-bent biases.

Unstructured data is the nucleus of an organization, irrespective of its size. The vast majority of an organization’s data is unstructured, thanks to organizations, which are heavily investing in IoT and embracing artificial intelligence. This kind of data holds huge potential for business leaders to leverage and gain an edge on initiating real-time campaigns through the use of content connectors. This type of data will help marketers understand their business results in more acute ways, anticipate and react quicker on risk and opportunity. Ged Parton, CEO of Maru Group, comments, “Text analytics technology is a vital resource to analyzing the thousands of unstructured data points generated every day through customer feedback, reviews and service interactions. But too often, as researchers, we’re not utilizing these tools to their full potential, clouding the technology by creating our own structures and code-frames and, in essence, introducing our own human biases into the results.”

Most leaders in digital marketing currently work with structured data, which is essentially organized. By applying scientific methodologies, we extrapolate hypotheses from this structured data and begin to formulate a predetermined mindset to create specific campaigns in various performance-based channels.

But what do we do with all of the unstructured data? Essentially data that hasn’t found a home? Unstructured data comes in many different formats, but the most established types of unstructured data may include books, journals, documents, metadata, health records, audio files, video, analog data, images, files, and unstructured text such as the body of an email message, a web page, or PDF document that languishes on your web properties. This unstructured data can reveal important and specific insights into your site visitors’ behavioral components, which can then be used to transform and deploy behavioral-based campaigns in real-time for optimal return. For example, we learned that IBM Watson is capable of understanding emotions when listening to certain audio files. This is widely known. Watson also is able to track and analyze inbound phone calls and extrapolate visitor sentiment. Harnessing sentiment is critical to developing these real-time campaigns.

Unfortunately, humans all too often allow unstructured data to permeate, so through no fault of their own, form unsuspected biases toward a visitor’s intent. This is the most critical common mistake humans do with AI: They form biases of this unstructured data, and these biases constrain the ability of machine learning to justify its full output potential. In a subsequent article, I’ll attempt to delve into ethics and biases regarding AI, which is tantamount to current projects that will change society. Still, today I want to focus on amplifying advanced digital marketing concepts using unstructured data.

Let’s just focus on content for a moment. Another way the subclasses of AI are extensively used in unstructured data is finding and closing gaps in your content marketing. Many modeling tools use machine learning to automatically and accurately group relevant comments together into clusters without the need for user-defined rules and human biases. It empowers users and researchers to identify key themes, understand relationships between trends, and uncover hidden patterns in data in just minutes.

Content marketing is “earned media,” and by narrowing this gap, you capture probably the most important segment of visitors you wouldn’t normally have found with a human brain. The essential narrative here is that performance metrics in content marketing are perceived to be more desirable from a “lead” perspective than paid media. This is why unstructured data is such a critical component for your marketing arsenal to discover gaps in content and deliver on higher than expected ROI.

From a branding perspective, monitoring unstructured data from various performance channels is also a critical step in your digital transformation journey, since it captures your users’ sentiment. The ability to harness this sentiment as a vacuum from disparate parts of the Internet is a game-changer to the health and wellness of a brand.

By immediately capturing sentiment, digital marketers instantly optimize channels and online web properties. Digital Marketers need to be nimble enough to capitalize on the positive sentiment and mitigate liability in case the sentiment turns negative. Machine learning and deep learning evaluate these brand mentions and immediately provide a score sentiment. This is obvious. By having AI monitor your brand in real-time, you can scale digital advertising expenditures to optimize performance when sentiment is positive. Further, you can improve individualized content experiences across all of your channels, assess and evolve creatives, and split test your landing pages across a bevy of tools. Other optimizations at scale include: moderating comments, monitoring ratings and reviews, and pushing geo-targeted ads through your mobile devices where sentiment is most dense.

Machine learning also determines which offers intrinsically motivate site visitors into action and engage users in conversations with bots. My favorite example of these multiple-choice bots for answers using AI is through a company called: www.adt.com Check them out. Their chat extension is highly engaging.

From a production perspective, unstructured data can be analyzed for grammar, sentiment, tone, and style. It can deliver on data-driven content and curate content from multiple sources, as mentioned in the example above. Unstructured data can build landing pages, develop real-time ad copy for your paid channels, and optimize site and web-property content for enhanced search results, speech, and text. Finally, this unstructured data can be translated into native languages while writing your next email subject lines.

I can’t say enough about how valuable your unstructured data is. If you are a brand that wants to execute on your digital transformation experience, harnessing your unstructured data should become a priority. When transforming your unstructured data, one thing to keep in mind is that being transparent with the site visitor is of utmost importance. Let the visitor know they are interacting with a chatbot or other AI-enabled features on your site, and this will alleviate any additional friction that may ensue.