Artificial Intelligence in Digital Marketing: Real Examples & Backed Future Predictions
Artificial intelligence (AI) is a fast-evolving technology and definitively a hot conversation topic. For some, AI is super scary, while for others, it is truly fascinating. Portraying AI as something evil that will destroy humanity might sell in Hollywood. I, however, think AI will be as moral as we build it. That is why creating a thinking machine that will use its intelligence to learn our human values will be critically important to preserve our safety.
So how can AI impact the future of digital marketing?
Hopefully, the following words will help you think about how the role of artificial intelligence will play in the future of digital marketing. Most importantly it should help you start navigating through the a.i. buzz so that you can be the driving force for innovation in your business and community.
Also, a better understanding of how a.i. can be used in digital marketing will help you to stay ahead and position yourself, your team and business for success.
Since technology is evolving extremely fast, it is my commitment to you as a reader to update this post every quarter, so you can always come back for fresh and updated information.
This article contains a lot of practical information. In order to make it easy to navigate, click on the following links, which will direct you to the part that you are most interested.
Understanding AI Better
- Understanding AI
- AI terminology you must understand
- How does artificial intelligence work
- Different types of AI
- The main AI Players
AI Examples in Marketing
- Content Marketing
- Search Engine Optimization (SEO)
- Customer Intelligence
- Website Development
- Customer Experience
- Marketing Automation
- Marketing Communications
Are we Going to get Replaced by AI?
Amazing facts about artificial intelligence and marketing that you may not know:
Click-though these high-level paradigm changes that occurred in past 20 years.
Understanding AI Better
If your grandparents ask you what is artificial intelligence, what would you answer?
There is no doubt that AI is a fascinating driving force for our future. Arguably, AI can power anything that is built on a line of code.
To offer some perspective, I like the way Andrew Ng, casts artificial intelligence as “the new electricity”.
Similar to the way in which electricity changed our world, a.i. will have a large impact on our society.
The important distinction is that i.a. will make computer machines more efficient and smarter but will not necessarily replace them. Before the birth of electricity, people used candles and natural light. When electricity became accessible, it improved the ability to have better, more efficient light, but it did not replace candles or a natural light. (impact on routine i.e. work during ‘light’ hours; electricity = new routine)
Making predictions for the future of machine intelligence can get controversial because many predictions, like flying cars, are far from becoming commercialized. Although there are visionaries like Mr. Musk who do inspire us to believe anything is possible.
Take a look at Elon Musk’s boring plan to solve traffic conjunction in Los Angeles.
AI Terminology you Must Understand
What is artificial intelligence?
Artificial intelligence is a broad field in computer science that simulates aspects of human intelligence and has the ability to interact with both virtual and real environments. Well-known areas of a.i. research include computer gaming, understanding social semantics and human language, robotics and sensory vision.
What is machine learning?
Machine learning refers to a process of using algorithms to accomplish a specific task. Technically speaking, it is the process of making improvements to a system’s performance through the use of data computations. For example, in marketing, you might be able to go through big data sets to identify the specific buying behavior patterns of your customers or to categorize the most frequent topics of your customers’ complaints.
What is deep machine learning?
Deep machine learning is a subfield of machine learning that essentially mirrors the structure and functions of our brain called the neural network. You can imagine deep machine learning as machine learning that has many layers due to complexity and extremely large data sets.
What is statistical learning?
Statistical learning uses data to form predictions and probabilistic models by complex statistical calculations. Statistical data with similar properties such as text fields, images or video content have a particular statistical regularity. That is why they can be calculated using probabilistic methods which can accurately process large amounts of data. In marketing, this can be us
ed to form models of your customers’ demographics and psycho-graphics.
What is cognitive computing?
For humans, learning is a human cognitive functions, similar to reasoning, creativity etc. . Cognitive computing is designed to have the ability to use “creative” reasoning about data, models, patterns and predetermined situations. You can think about cognitive computing as an ecosystem that uses both machine learning and statistical learning algorithms to function independently.
How does Artificial Intelligence Work?
If you think about it, computers are still pretty dumb because they do exactly what we tell them to do and they are infants when it comes to learning.
That’s why AI is an extremely hard problem. The other part is that machine learning depends on clean data. This is a big challenge because a.i. can only be as good as the quality of data available.
The objective of AI is to enable machines to learn and think like humans.
Today, AI uses algorithms to form models and generate insights from data in order to make decisions and generate predictions.
Credit to NVidia
Developing artificial intelligence for digital marketing purposes does not dramatically differ from developing a.i. for self-driving cars or other systems like Spotify or Netflix. You probably noticed that both Spotify and Netflix use algorithms to recommend music and movies to you.
In this example, machine learning is just trying to make your life easier. By collecting data from the choices you made in past, it can offer you recommendations for music and movies that might spark your interest in the future. The more data the algorithm can collect, the more it can learn about your personal preferences and tailor recommendations to your taste.
There are Different Types of AI
My goal here is to explain the different types of AI as clear as possible, so that even if you are a nontechnical person, you can still grasp it.
- The most simple form of a.i. is created with simple conditional rules (i.e. if this, then that). Think of using Zapier or ITTT to create sequences to automate your marketing processes.
- The more complex a.i. model would still use conditional rules but only in a particular environment or circumstance. Smart robot vacuums are a good example of this model. They are programmed to know in which room (environment) they should clean dirt and dust.
- An even more complex AI model would be programmed to accomplish a particular goal. That’s where statistical or machine learning would try to create a model based on your provided data.
- The highest level of a.i. would have the sophistication to learn itself, from accessing large data sets. Such agent would use what I described before as cognitive learning: combining statistical learning and deep machine learning to “reason”. Creating self-driving cars is a great example of this model.
In order to get a self-driving car on the road, there are multiple subtle processes (traffic rules, environmental monitoring scans, car performance, etc.) that are transformed into a series of algorithms and computations that makes self-driving cars a reality.
The Main AI Players
When it comes to artificial intelligence, I believe there are three different types of companies focusing on AI development: the giants, the service providers, and the innovators.
- The giants: Big companies like Google, Amazon, Microsoft, Facebook, Baidu, Shopify and Tencent all made significant investments in AI development. Most importantly, they power AI with their own data. These companies have an enormous advantage because they have access to massive amounts of well-structured data. Even though data for a.i. development has become more democratized, the giants control the context of a.i. development because they own most of the data.
- The service providers: These are companies like Salesforce, Oracle, Mitel or smaller consultancy companies. Their role is to help other companies and governments make practical use of big data. Typically, they are able to apply statistical learning to develop insights and models to evaluate big data sets.
- The innovators: These are companies like DeepMind, Sizmek, Cruise Automation (acquired by General Motors), Bosh or Zebra that are not focused on servicing and don’t own data. Rather, they use 3rd party data to power their AI. These companies strive to solve big problems like self-driving cars, medical research, programmatic media placement or flexible airfare pricing.
Examples of AI & Machine Learning in Marketing
Technology is moving fast. With the progression of machine learning, the world around us will feel faster and human convenience will reach new heights.
A study about generation Z from the J. Walter Thompson research firm will give you a snap shot of how millennials think about AI in a retail setting. 70% of respondents indicated that they would like retailers to use AI to show more interesting and tailored products.
Eric Schmidt from Google recently said that AI will get companies to IPO (initial public offering) faster.
Examples in Content Marketing
Many of us benefit from content personalization on daily basis.
If you are watching Netflix or listening to music on Spotify, you may notice that you are given a recommendation based on your taste. Both Netflix and Spotify use a.i. to collect data about your past choices to better understand what you may like in future. The more data they can collect, the more they can learn about your preferences.
You can deploy content personalization machine learning if you have a large library of content like blogs, on-demand webinars, videos or infographics. Assuming that you have a fair amount of visitors returning to your website, you can satisfy your visitors by offering personalized content.
If you are interested getting insights into your content ask AI agent Alfred. Alfred is being developed by Boomtrain and it can serve as your content quality advisor.
Content generation intelligence may sound like a threat for content creators. In reality, it powers them to create better content. Imagine you can spend less time on reporting or preparing business summaries for your supervisor. Rather that burdening yourself with this administrative exercise, you can focus on the creative aspects that will make your content stand out.
Currently, AI that focuses on generating content uses structured data to segment related information into categories. For example, this AI model will be able to automatically create reports or summaries for briefings and news articles.
I’ve seen a few attempts of AI agents (like Wordsmith platform) that try to leverage unstructured data from search engines to create content. Luckily for content writers today, this intelligence is still under development.
Examples in Advertising
In past years, programmatic advertising has garnered a negative reputation. As a groundbreaking technology, it made big promises for effective advertising. A lot of money was spent, however with little return. With more data available and the evolution of AI technology, this is changing.
According to Google, over 80% of online advertising will be programmatic by 2018.
Essentially, programmatic computing works like a recommendation engine for advanced ad targeting. Optimizing ad campaign performance requires a lot of data analysis, which machines can do more quickly and accurately than humans. Analyzing top performing ad copies and benchmarking them against specific audience demographics is a variable that can be programmed. Although algorithms can significantly improve the performance of media buying, I don’t think that machine learning will replace “the creative” part of advertising anytime soon.
Examples in Search Engine Optimization (SEO)
Intelligent Search engines
Look at the Google RankBrain algorithm.
When you start searching your target queries, RankBrain interprets very large amounts of indexed data to give you the most relevant results. The goal is to provide instant and user-centered answers to any questions you have.
If you remember a time when part of SEO involved adding meta tags on pages, that time is now over. RankBrain is much smarter as it can comprehend the meaning of text, audio and video. That is why SEO agencies are placing more emphasis on amazing and quality content rather than plain keyword matching and title optimization.
Most companies love SEO as it allows them to acquire customers at a low cost.
Since tools are a vital necessity for any SEO specialist, here is list of a few AI powered tools.
- Ubbersuggest.io (recently acquired by Neil Patel) as well as Answerthepublic helps content creators to know what keyword phrases to include to achieve a better topical relevancy.
- MarketMuse helps with semantic optimization by measuring the quality of your content.
- Ahrefs & SEMRush crawl the web to help you get more information about your competitors keywords, backlinks, ads and more.
Examples in Customer Intelligence
Social media provides a fantastic opportunity for deep machine learning.
Think about this.
Facebook is currently the social media leader when it comes to number of users – it just surpassed 2 billion. Think about the gigantic amount of data Facebook is collecting. Anything from your activity, information about your past, the location where you currently live, your interests, friends and family, your political preference, language or emotional reactions to specific types of content.
A bit scary, isn’t it?
For machine learning, these large data sets present a fantastic opportunity to better understand not only individuals, but different ethnic groups, nations and society at large. Anything from users’ purchasing behaviours, product recommendations to sentiment and mood analysis can be identified by Facebook’s AI.
This chart displays the number (in millions) of users per social media platform.
Because Facebook is collecting enormous amount of data, Facebook Ad Manager is becoming a super powerful tool.
There are many benefits for marketers. These include precise ad targeting, advanced bidding strategies to capture purchase micro-moments and detailed customer segmentation capability. You can use these for your lead generation, brand awareness and customer retention programs.
With respect to customer segmentation intelligence, take a look at predictive customer data hub AgilOne. A CRM ( customer relationship manager ) focusing on a deep understanding of your customers’ behaviours, trends and revenue opportunities as opposed to serving only as a customer database.
Companies with large product catalogs, international customer base, and multiple currency offerings can significantly benefit from dynamic pricing. Price is always relative to the market competition, your product positioning and, to a certain extent, what your customer is willing to pay.
By using dynamic pricing, you will create the perception of a larger offering. It is also a strategic way of maximizing your profit margin.
A.i. allows big market places like Ebay or Amazon to incorporate dynamic pricing which can drive competitiveness and earn customer business as they can offer better prices.
Aligning all these factors by using an algorithm is a huge time-saver for analyzing correlations between market pricing trends, sales demand, capacity and product inventory levels.
Airline companies are an obvious example. As a customer, you can find multiple prices for a flight from Miami to New York. All of these price differences are managed by an algorithm that takes into account factors such as base fare, taxes and airport fees, fuel surcharge, service fees, class preference, seat preference, food and drinks, etc.
The image below is an example of the Axa insurance machine learning pricing model.
Examples in Website Development
Adaptive Website Design
A.i. is starting to play a big role in website development.
It won’t replace web designers but it will enable them to produce work more quickly and power them with more design and development capabilities.
The Grid is a website design platform powered by an a.i. named Molly. Molly is able to crop and recognize images, select a proper font and use its algorithmic palette to assemble website layouts and customize content in a visually appealing manner. Play this video to see how to build AI powered website.
Examples in Customer Experience
Predictive customer service
Satisfying customers’ requests quickly and elegantly is the key to customer happiness.
Good communication is an important avenue for making that happen. You can use a.i. to research large data sets and analyze hundreds of different communication patterns, behaviours and probability scenarios. This can dramatically improve the likelihood of treating each customer in the exact way he or she wishes.
Most of the aspects of customer service will be available 24/7. There will be very little waiting time on phone lines or in live chats. Take a look at the Growthbot movement – a chatbot for marketing and sales.
Saffron is a neural intelligence platform that analyzes large sets of data to understand how people make decisions and what information and associations matter to them. Play this video for quick overview.
Integration between offline and online
Businesses with both a physical location and online presence will be able to leverage machine learning to combine the online and offline experience for their shoppers. In the future, we will see more business models like Amazon Go or Amazon Fresh Pick Up. That is because companies will be able to use AI for both security and consumer convenience, providing a sense of autonomy for the shopper.
Examples in Marketing Automation
To me, time is a valuable commodity. That is why I can’t image not using marketing automation shortcuts to reduce the burden of administrative tasks.
The game is getting even better.
What if you have an AI agent that can monitor your daily workflows and recommend a predictive solution to your current processes?
Building a strong demand generation engine is challenging and time-consuming. In the next few years, sales and marketing will be tremendously accelerated by autonomous automation, predictive lead scoring, email marketing match-making and content recommendations for buyers journeys. Acquiring high-quality leads and building a predictive sales pipeline will be a smooth walk.
If you are looking to deploy predictive lead scoring, I am personally familiar with Hubspot but there are other solutions like:
- Salesforce Einstein – AI designed to power anything from sales & marketing to community mangement and commerce.
- Mintigo – A predictive intelligence platform for sales and marketing.
- Infer – Platform for driving predictive customer acquisition.
- Wise.io – Predictive lead scoring developed under the umbrella of GE Digital.
Email marketing match-making will be similar to predictive customer service. The performance of email marketing campaigns and funnels will be automatically analyzed and paired with other customer intelligence data to produce recommendations for future campaigns that will be programmed to accomplish a specific goal.
Examples in Marketing Communications
Chatbots are the modern communicators. A bot algorithm like The Kit (acquired by Shopify) can communicate or perform a set of specific actions.
If you had the chance to watch the Facebook F8 conference, you may know how chatbots are currently being tested to replace traditional 1-800 numbers. At Growth Media, we love using bots for social networks and different communication channels. If you want to connect with us, now you have the option to connect with one click on Facebook Messenger.
The use of personal assistants is on the rise.
Many people use voice recognition agents on a daily basis. Translation of speech progressed significantly due to the development of neural networks.
Speech recognition agents like Siri (Apple), Cortana (Microsoft), Alexa (Amazon) or M (Facebook) are becoming more popular. While most consumers think about them only as gadgets, these AI agents collect super powerful information about personal consumer habits from any aspect of your life. In another words, these AI agents have the power to control context for communicating information you don’t know.
Speech recognition is booming in the Chinese market, where people prefer talking vs. using a keyboard to type small and intricate characters. Baidu is making big strides on this front with voice search.
While most personal assistants (PA) are under constant development, there are some notable skills which the agents can already execute. By understanding human language, PAs can give you information about news, flight reservations, weather, calendar appointments, music, directions, movies and small businesses. Siri and Alexa integrate with home appliances, including pet feeders, and have an open API ( application programming interface) for 3rd party app developers.
The long term goal for personal assistants is to have the capacity to accomplish anything humans will ask for. All personal assistants are fast, providing people with a high level of convenience. It appears like there will be very little navigation required as most of the tech gadgets, cars and home automation will be done through asking questions and giving orders.
Are we Going to get Replaced by AI?
This is the million dollar question and is subject to frequent debate.
I do not think that AI will replace marketers anytime soon. Artificial intelligence is still at the very beginning of prototyping, fixing and testing all errors. On the flip side, I think marketers should start actively using and testing AI that is available in order to experience it and learn to collaborate with it.
Meanwhile, the AI of today will help marketers to skip the boring stuff and hopefully make them more focused and smarter. If you have a talented marketing team and the resources to play with intelligent tools, you will get the best results.
However, in the future we might see particular roles being replaced. The first people to be let go will likely be those roles that are very task-based and lack creativity.
The future of artificial intelligence in marketing is unpredictable. What is predictable is the intention for AI development. Five years from now, we might see amazing game-changing inventions as well as continuous steady progress towards more intelligent, self-thinking computers.
To close this article, watch this powerful Ted talk from Nick Bostrom. He shares a thought-provoking perspective about what happens when our computers will get smarter then we are. Enjoy!
50 Actionable Tips for Growth Marketers I created this blog to share some growth marketing tips you can utilize today to help with their operations….