Making Sense of AI: What Adobe Sensei Means for You
Artificial intelligence is transforming creativity and business — and Adobe Sensei is providing the technology to do it.
There’s an inevitability that comes with any “cutting-edge” technology — within a few years or, even, months, that once envelope-pushing advancement is going to be seen as completely commonplace. We’ve seen this for decades, especially when it comes to defining and framing artificial intelligence (AI).
As Wired’s founding editor Kevin Kelley once explained, this isn’t a new phenomenon. In the 1960s, programming was considered AI. Fast forward to the 1980s, and programming was the norm while databases and more modernized approaches to statistics were considered AI. Soon enough, though, these former advances became par for the course — basic technological applications that didn’t warrant any sort of fanfare, let alone the AI title. “Every achievement in artificial intelligence redefines that success as ‘not AI,’” he wrote.
What artificial intelligence looks like today
AI today continues to shift and evolve. However, there does seem to be more of a popular consensus when it comes to defining present-day AI. Very simply, most experts see AI as programming computers so they’re able to perform tasks previously done by humans. Under this umbrella, AI enables the computer to “reason” so it can identify insights and make predictions based on complicated information.
At Adobe, we have a similar anchor for AI, looking at advances with a three-part criteria that looks at the technology used, the value it delivers to customers, and the alignment with our corporate values. Simply put —
Adobe Sensei delivers the magic of artificial intelligence and machine-learning technology by enabling customers to:
- Discover what is hidden,
- Accelerate what is slow,
- Decide when it matters
while it serves the creator and respects the consumer.
In-line with this definition, we introduced Adobe Sensei in October 2016. Adobe Sensei brings together two unique Adobe capabilities — a massive volume of content and data, plus deep understanding of how customers work.
This paired with the latest advancements in AI, machine learning, deep learning, and related fields enable customers to discover what is hidden, accelerate what is slow, and decide when it matters. All of this happens while still aligning to our corporate values of serving the creator and respecting the consumer by checking these three boxes:
#1. Delivering artificial intelligence and machine learning at its core
Science fiction writer Arthur C. Clarke was clearly peering into the future when he wrote, “Any sufficiently advanced technology is indistinguishable from magic.”
Ultimately, we want to bring the magic of technology to the user experience. To do this, Adobe Sensei integrates all branches of artificial intelligence, including machine learning or deep learning. This enables complex problems and processes to be broken down into much simpler terms so better solutions can be evaluated and acted on.
Given Adobe Sensei’s scale and scope, this technology does tasks that seem unreal — magic, almost. From predicting customer churn to adding a sunset to the background image of a photo, Adobe Sensei leverages machine learning, algorithms, and outputs to improve processes without requiring a human programmer intervene. The more data received, the better the performance.
#2. Promoting customer-first value
Adobe Research is tasked with imagining and inventing the future, with an eye on experimentation and collaboration. Last year, the team published more than 170 technical papers, filed for more than 130 patents, and transferred nearly 60 new technologies to existing Adobe products.
With everything we do, we’re laser-focused on adding “magic” to our collective toolbox, ensuring we can create a better, more sophisticated, and more high-value future for our customers and their customers. Lately this has included AI and machine-learning capabilities, which help our customers deliver customer-first experiences with greater speed and relevance than ever before.
That said, not all AI syncs with Adobe Sensei’s goals and value-centric approach. For us to incorporate new AI and machine-learning extensions, they must provide clear-cut customer value in at least one of three ways:
Discover what is hidden
A 2018 Gallup survey found that, despite living in a period of information overload, the explosion of information makes most people feel it is more difficult to stay “in the know.” In the same way, marketers have access to so much information about customers that it can be harder to understand those customers — it’s difficult to separate the important findings from the noise.
Adobe Sensei can help create better, more effective, and engaging experiences by discovering the hidden jewels in the great mounds of data. Anomaly Detection in Adobe Analytics, for example, uses machine learning to assess and analyze customer behavior. By monitoring past behavior and comparing findings to real-time patterns, Anomaly Detection can spot behavioral outliers, separating the norm from “the noise.”
Accelerate what is slow
Got an idea for an image for a marketing campaign — an idea that’s so powerful you can literally see it in your head? Finding the right asset can be a time-consuming and, often, tedious chore.
To simplify the process, Adobe Sensei powers the Smart Tags feature in Adobe Experience Manager. By automatically adding metadata tags to images, Smart Tags make it easy to find the right images even within bulk files.
That’s just the beginning. To help increase efficiency and effectiveness, Adobe Sensei’s “smart cropping” feature automatically crops images to specific dimensions while keeping the salient object in the picture.
These are both examples of activities humans could do manually with considerable time and effort. What’s more, these tasks are challenging if not impossible to do at the scale when necessary. Adobe Sensei tackles these smart processes — and takes these critical steps to get the job done — automatically.
Decide when it matters
Adobe Sensei makes decisions at a speed and scale that are otherwise impossible to do manually. By turning over essential tasks like Automated Personalization in Adobe Target, Adobe Sensei can analyze patterns and end results, optimizing experiences for every customer and segment. In looking at everything — real-time decisions, customer traits, temporal considerations, and more — Adobe Sensei can optimize experiences in the moment and over time.
#3. Aligning with Adobe’s core values and customer-first commitment
From an AI perspective, our overriding philosophy is to “serve the creator and respect the consumer” — and, leveraged properly, it does:
- Think about those repetitive tasks that can now be handled by a machine marketing partner.
- Think about the enhanced data capabilities and the better customer experiences now available.
- Think about the improved human-machine interactions that enable greater productivity.
A good example is Predictive Fatigue Management, an upcoming feature in Adobe Campaign, which applies email frequency controls to limit the number of emails a customer receives. Reducing the number of unwanted emails serves both the customer and marketer, eliminating irritation for the recipients while boosting the conversion rate. Reducing spam merges business benefits with data ethics.
And that intersection is at the heart of how we view artificial intelligence. Used effectively, it can provide huge productivity benefits for creative and marketers, giving them the time and headspace to find the next big idea.
Like any powerful technology, though, ours must be used responsibly. We at Adobe remain committed to serving the needs of customers and consumers in an ethical manner, in alignment with our own core values.
Adobe Sensei is a powerful tool — and, like any powerful technology, we must always be mindful of the implications and keenly aware of how they impact the customer. For this reason, we’ve formed a working group to examine the impact of technology, evaluating and creating best practices. This scope includes data privacy, data governance, and data diversity — making sure data used to develop AI-driven algorithms doesn’t inadvertently produce biased algorithms.
The Adobe Research team recently trained a deep learning neural network to identify image manipulation — something forensic experts used to devote hours to, but now AI technology can identify in seconds. This, among other strategies, can help curb everything from image tampering to image noise, ensuring greater authenticity in creative and content.
“It’s important to develop technology responsibly, but ultimately these technologies are created in service to society,” says Jon Brandt, senior principal scientist and director for Adobe Research. “Consequently, we all share the responsibility to address potential negative impacts of new technologies through changes to our social institutions and conventions.”