Adobe’s New Virtual Analyst Answers Questions You Should Be Asking


Artificial intelligence has been promising to revolutionize marketing

for some time now.

Indeed, it feels as if new martech tools and technologies have long been on the cusp of taking those aspects of marketing that use speech and image recognition, machine learning and semantic searching, for example, to the next level.

In fact, such developments have led most industry professionals to expect AI to introduce a new suite of self-driving marketing automation – powered especially by machine learning – that will transform sales and marketing.

The idea is simple: New software will use machine learning to perform marketing operations without any explicit input from marketers, offering opportunities in predictive lead generation, content recommendations and even e-mail acquisitions, among other channels.

Adobe seems to be making all this real with the introduction of a new “virtual analyst” to its Adobe Analytics platform. The new robot assistant is laying the groundwork for a new layer of data analytics.

Most existing martech tools use various technologies to provide useful information collected from large swaths of data. The tools and services analyze big data to produce valuable marketing insights that can be used, for example, to improve the customer experience, manage customer relationships or even deliver revenue growth and cost reductions.

Crucially, however, to achieve any of these outcomes, these AI-powered technologies require human input and maintenance to know what to do with the data accumulated in the first place. According to Adobe, marketing departments fail to incorporate as much as 97-99% of the data gathered because they lack the resources to handle the staggering volume.

Think of it this way: Although marketers may have access to massive amounts of data, the value of that information is defined by the questions asked and the specific alerts programmed manually into AI-powered martech tools.

What, then, of the many questions that aren’t asked? More importantly, what about the questions that marketers won’t even know they should be asking?

What should you be asking a virtual analyst

Adobe’s virtual analyst flips this human-robot relationship on its head, using AI to provide entirely new sets of data insights in a hyper-competitive marketing landscape – one where marketers need regular, in-depth and customized information based on real-time analyticsOpens a new window in order to personalize messages and products delivered to customers.

This new AI-powered assistant uses Adobe’s AI and machine learning model, Adobe Sensei, to discover what the company refers to as “unknown unknowns.” It looks for rare but significant patterns in user and customer data to pinpoint marketing opportunities that Adobe’s marketing clients’ didn’t even think to look for.

“This can include valuable insights around what drove unexpected spikes or drops in key metrics,” Adobe claims, “such as online orders and web traffic, critical to addressing escalations in real-time and finding new opportunities to grow revenue.”

In the testing phase of the cognitive analyst, one Adobe customer found out its content was being stolen by a competitor – an unexpected discovery uncovered by the robot assistant as it dug into specific spikes to sections of the customer’s website.

What makes the new AI especially unique is that it can adapt what it looks for based on how various people within a company use it. The different questions they ask about the same data will help guide the robot in its analysis of irregularities it uncovers in the data.

Users can give “like” and “don’t like” feedback to different findings, which will help the analyst determine and prioritize the most valuable types of insights within the context of the marketer’s company, offering clients an increasingly personalized product that uses real-time analytics.

It can then automatically unearth valuable information without explicitly being ordered, and without the need for manual inputs from marketers.

Ideally, Adobe’s end game is for the tool to be smart enough in its reporting so that in-depth knowledge of data science will be unnecessary to understand the findings.

Data management is now an integral piece of a company’s marketing stack, and finding ways to manage the data overload has become pivotal. Adobe’s new tool is one of the first on the market that promises to improve the user experience – and offer a competitive advantage.

Watch this space, because you can bet plenty more of these virtual assistants are on their way, and we’ll keep an eye out for them.