*Update: To watch a presentation given on this topic by the author, click here.
The artificial intelligence revolution has arrived. New AI models are being released at lightning speed, and they are drastically changing the way we interact with the digital world.Below are some incredible recent advances:
- DALL-E 2 and Mid Journey: Generate compelling images from a prompt, to provide unique imagery that speaks directly to the user and their specific need.
- ChatGPT, LaMBDA, and LLaMA: Return text completion in natural language from any text prompt such as a phrase or a sentence. Can generate convincing human-level essays, outlines, or marketing copy.
- Codex: Turns natural language prompts into coding suggestions across dozens of programming languages.
This level of customization won’t be limited to generating pictures, copy or code from text. Its capabilities will span to movies, TV and music, even generating totally new styles of by mingling a person's preferences into an AI musician. Someday, It may be possible to integrate various AI features such as these together to serve a truly unique experience in real time. Although the timing is difficult to predict, one thing for certain is that change is coming to digital experience creation. The future is exciting, but what can be acted upon today?
First, some history. At the advent of the world wide web, browsers served static pages with links, pictures and text that were the same for everyone. There was no way for website owners to identify visitors’ individual preferences. Over time, they gained the ability to do so to a limited extent, only via account creation. Cookies followed, which stored a little more info without requiring accounts, and could be used to interpret a customer's goal to serve them better content.
However, cookies have limitations. At the start of the first visit, a cookie has little to no details, so it cannot be used to select personalized user experiences. This limitation was insurmountable until recently. Now it is possible to predict a user's goal without requiring information stored in cookies by analyzing non-identifying details such as the time of day someone visits, their browser type, or their operating system. Combining these system-level details and using AI to analyze their relationship to personalized content can unveil visitor demographics previously impossible to determine. This novel method of personalization serves tailored content for each visitor to an unprecedented degree.
Today, an artificial intelligence feature within Adobe Target uses this method to select from among numerous versions of a website to serve personalized user experiences in real time. Its machine learning algorithm is enterprise-ready, and can analyze numerous data points to select and serve unique web pages at any scale. These capabilities showcase how years of machine learning advances are incorporated into Adobe products from Photoshop to the Experience Cloud.
There are a myriad of data points that can used to identify a demographic. Some of the details analyzed may not seem meaningful on the surface. However, it may turn out that customers visiting the website on Tuesdays at 8 AM from California using the Chrome browser, with low screen resolution are, for whatever reason, likelier to be attempting to pay their monthly bill. A person would be unable to predicting these specific behaviors across large data sets, but AI models can. AI has the ability to identify patterns in customer behavior to make predictions. As more customers fulfill their needs, it learns and finds correlations between visitor information and their completed goals. The next time someone visits the website with similar characteristics, the model can predict their desired goal with impressively high accuracy.
If someone is visiting their cell phone service provider’s website to pay their bill, the AI predicting this at the beginning of the session might make the first button displayed "Pay Your Bill". It then takes the customer directly to a page that is normally four clicks away, buried in account settings. With this level of personalization, enterprises can make each customer's life easier by minimizing the number of steps needed to complete their goal. The scale at which AI can accomplish this for enterprise is expansive, and real-time data helps train the model as it evaluates more and more visitor traffic.
There are privacy concerns, as some consumers do not want their data tracked and would therefore prefer not to receive personalized experiences. With these customers, and individual privacy rights in mind, Adobe Target offers options for opt-out. However, Adobe’s AI personalization capabilities ultimately provide customers with the digital experiences they want. Research from McKinsey shows that 71% of those polled respond positively to personalization, because it addresses their specific needs, whether that be getting a product recommendation or providing a clear path to changing a particular account setting. Benefits must be carefully balanced with potential privacy related drawbacks, and precautions must be taken in consideration of these competing perspectives.
Adobe takes privacy seriously, advocating for ethical considerations in the development of AI systems. There are means of designing these models to process anonymized data along with terms of service and internal policies that protect users. To accomplish this, Adobe institutes strict processes in configuring how proprietary AI models learn. Additionally, website owners have the ability to control and limit the data to be leveraged via robust data filters. With these filters, Target provides controls to define the exact attributes to be considered by the model that are most important to the business case, and excludes those which are not. This ability to customize AI models and control which attributes are most important can deliver faster results and ultimately provide higher conversions on specific business metrics.
Adobe Target’s AI capabilities make tailored experiences possible for each user’s unique preferences and needs. This saves customers time and frustration, leading to richer experiences and more successful shopping sessions.
This kind of personalization is just the beginning of what could be. Serving everything from creative design to product merchandising, it’s fun to imagine future forms of generative AI crafting experiences to meet the needs of a user's unique situation. But today, we have a system that can identify the perfect copy, creative imagery, and call to action to provide a richer experience for the customer.
The natural extension is to combine various experiences into a unique story far more effective than its one-size-fits-all counterpart. This lets us create delightful digital journeys, subsequently improving conversion rates and generating higher revenues, while providing an experience that customers will love.
It’s not just a vision of the future. These systems exist today, and Adobe's technology can implement them at scale.
Learn how to utilize Adobe Target to deliver highly personalized experiences that lift conversion ra...
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.